Understanding the effects of mesenchymal stromal cell therapy for treating osteoarthritis using an in vitro co-culture model.

Osteoarthritis (OA) is a leading cause of chronic pain and disability, for which there is no cure. Mesenchymal stromal cells (MSCs) have been used in clinical trials for treating OA due to their unique ability to generate paracrine anti-inflammatory and trophic signals. Interestingly, these studies have shown mainly short-term effects of MSCs in improving pain and joint function, rather than sustained and consistent benefits. This may reflect a change or loss in the therapeutic effects of MSCs after intra-articular injection. The present study aimed to unravel the reasons behind the variable efficacy of MSC injections for OA using an in vitro co-culture model. Osteoarthritic human synovial fibroblasts (OA-HSFs) were co-cultured with MSCs to investigate their reciprocal effects on cell responses and whether a short-term exposure of OA cells to MSCs was sufficient for reducing their diseased characteristics in a sustained manner. Gene expression and histological analyses were performed. OA-HSFs exposed to MSCs showed short-term downregulation of inflammatory markers. However, the MSCs showed upregulation of inflammatory markers and impaired ability to undergo osteogenesis and chondrogenesis in the presence of OA-HSFs. Moreover, short-term exposure of OA-HSFs to MSCs was found to be insufficient for inducing sustained changes to their diseased behaviour. These findings suggested that MSCs may not provide long-term effects in correcting the OA joint environment due to them adopting the diseased phenotype of the surrounding tissues, which has important implications for the future development of effective stem-cell-based OA treatments with long-term therapeutic efficacy.

OA foundations - experimental models of osteoarthritis.

Osteoarthritis (OA) is increasingly recognised as a disease of diverse phenotypes with variable clinical presentation, progression, and response to therapeutic intervention. This same diversity is readily apparent in the many animal models of OA. However, model selection, study design, and interpretation of resultant findings, are not routinely done in the context of the target human (or veterinary) patient OA sub-population or phenotype. This review discusses the selection and use of animal models of OA in discovery and therapeutic-development research. Beyond evaluation of the different animal models on offer, this review suggests focussing the approach to OA-animal model selection on study objective(s), alignment of available models with OA-patient sub-types, and the resources available to achieve valid and translatable results. How this approach impacts model selection is discussed and an experimental design checklist for selecting the optimal model(s) is proposed. This approach should act as a guide to new researchers and a reminder to those already in the field, as to issues that need to be considered before embarking on in vivo pre-clinical research. The ultimate purpose of using an OA animal model is to provide the best possible evidence if, how, when and where a molecule, pathway, cell or process is important in clinical disease. By definition this requires both model and study outcomes to align with and be predictive of outcomes in patients. Keeping this at the forefront of research using pre-clinical OA models, will go a long way to improving the quality of evidence and its translational value.

Metformin as a potential disease-modifying drug in osteoarthritis: a systematic review of pre-clinical and human studies.

OBJECTIVE: Osteoarthritis causes significant pain and disability with no approved disease-modifying drugs. We systematically reviewed the evidence from both pre-clinical and human studies for the potential disease-modifying effect of metformin in osteoarthritis. METHODS: Ovid Medline, Embase and CINAHL were searched between inception and June 2021 using MeSH terms and key words to identify studies examining the association between metformin use and outcome measures related to osteoarthritis. Two reviewers performed the risk of bias assessment and 3 reviewers extracted data independently. Qualitative evidence synthesis was performed. This systematic review is registered on PROSPERO (CRD42021261052 and CRD42021261060). RESULTS: Fifteen (10 pre-clinical and 5 human) studies were included. Most studies (10 pre-clinical and 3 human) assessed the effect of metformin using knee osteoarthritis models. In pre-clinical studies, metformin was assessed for the effect on structural outcomes (n = 10); immunomodulation (n = 5); pain (n = 4); and molecular pathways of its effect in osteoarthritis (n = 7). For human studies, metformin was evaluated for the effect on structural progression (n = 3); pain (n = 1); and immunomodulation (n = 1). Overall, pre-clinical studies consistently showed metformin having a chondroprotective, immunomodulatory and analgesic effect in osteoarthritis, predominantly mediated by adenosine monophosphate-activated protein kinase activation. Evidence from human studies, although limited, was consistent with findings in pre-clinical studies. CONCLUSION: We found consistent evidence across pre-clinical and human studies to support a favourable effect of metformin on chondroprotection, immunomodulation and pain reduction in knee osteoarthritis. Further high-quality clinical trials are needed to confirm these findings as metformin could be a novel therapeutic drug for the treatment of osteoarthritis.

Pharmacological characterization of GLPG1972/S201086, a potent and selective small-molecule inhibitor of ADAMTS5.

OBJECTIVE: A disintegrin and metalloproteinase with thrombospondin motifs 5 (ADAMTS5) is a key enzyme in degradation of cartilage in osteoarthritis (OA). We report the pharmacological characterization of GLPG1972/S201086, a new, potent and selective small-molecule ADAMTS5 inhibitor. METHODS: Potency and selectivity of GLPG1972/S201086 for ADAMTS5 were determined using fluorescently labeled peptide substrates. Inhibitory effects of GLPG1972/S201086 on interleukin-1α-stimulated glycosaminoglycan release in mouse femoral head cartilage explants and on interleukin-1ß-stimulated release of an ADAMTS5-derived aggrecan neoepitope (quantified with ELISA) in human articular cartilage explants were determined. In the destabilization of the medial meniscus (DMM) mouse and menisectomized (MNX) rat models, effects of oral GLPG1972/S201086 on relevant OA histological and histomorphometric parameters were evaluated. RESULTS: GLPG1972/S201086 inhibited human and rat ADAMTS5 (IC50 ± SD: 19 ± 2 nM and <23 ± 1 nM, respectively), with 8-fold selectivity over ADAMTS4, and 60->5,000-fold selectivity over other related proteases in humans. GLPG1972/S201086 dose-dependently inhibited cytokine-stimulated aggrenolysis in mouse and human cartilage explants (100% at 20 µM and 10 µM, respectively). In DMM mice, GLPG1972/S201086 (30-120 mg/kg b.i.d) vs vehicle reduced femorotibial cartilage proteoglycan loss (23-37%), cartilage structural damage (23-39%) and subchondral bone sclerosis (21-36%). In MNX rats, GLPG1972/S201086 (10-50 mg/kg b.i.d) vs vehicle reduced cartilage damage (OARSI score reduction, 6-23%), and decreased proteoglycan loss (∼27%) and subchondral bone sclerosis (77-110%). CONCLUSIONS: GLPG1972/S201086 is a potent, selective and orally available ADAMTS5 inhibitor, demonstrating significant protective efficacy on both cartilage and subchondral bone in two relevant in vivo preclinical OA models.

Pathology-pain relationships in different osteoarthritis animal model phenotypes: it matters what you measure, when you measure, and how you got there.

OBJECTIVE: To determine whether osteoarthritis (OA) pain characteristics and mechanistic pathways in pre-clinical models are phenotype-specific. DESIGN: Male 11-week-old C57BL6 mice had unilateral medial-meniscal-destabilization (DMM) or antigen-induced-arthritis (AIA), vs sham-surgery/immunised-controls (Sham/Im-CT). Pain behaviour (allodynia, mechanical- and thermal-hyperalgesia, hindlimb static weight-bearing, stride-length) and lumbar dorsal root ganglia (DRG) gene-expression were measured at baseline, day-3, week-1/-2/-4/-8/-16, and pain-behaviour:gene-expression:joint-pathology associations investigated. RESULTS: DMM and AIA induced structural OA defined by progressively increasing cartilage erosion, subchondral bone sclerosis and osteophyte size and maturation. All pain-behaviours were modified, with model-specific differences in severity and temporal pattern. Tactile allodynia developed acutely in both models and persisted to week-16. During early-OA (wk4-8) there was; reduced right hindlimb weight-bearing in AIA; thermal-hyperalgesia and reduced stride-length in DMM. During chronic-OA (wk12-16); mechanical-hyperalgesia and reduced right hindlimb weight-bearing were observed in DMM only. There were no associations in either model between different pain-behaviour outcomes. A coordinated DRG-expression profile was observed in sham and Im-CT for all 11 genes tested, but not in AIA and DMM. At wk-16 despite equivalent joint pathology, changes in DRG-expression (Calca, Trpa1, Trpv1, Trpv4) were observed only in DMM. In AIA mechanical-hyperalgesia was associated with Trpv1 (r = -0.79) and Il1b (r = 0.53). In DMM stride-length was associated with Calca, Tac1, Trpv1, Trpv2, Trpv4 and Adamts5 (r = 0.4-0.57). DRG gene-expression change was correlated with subchondral-bone sclerosis in DMM, and cartilage damage in AIA. Positive pain-behaviour:joint-pathology associations were only present in AIA - for synovitis, subchondral-bone resorption, chondrocyte-hypertrophy and cartilage damage. CONCLUSION: Pain and peripheral sensory neuronal responses are OA-phenotype-specific with distinct pathology:pain-outcome:molecular-mechanism relationships.

Differential patterns of pathology in and interaction between joint tissues in long-term osteoarthritis with different initiating causes: phenotype matters.

OBJECTIVE: To determine if osteoarthritis (OA) progression and joint tissue-pathology associations link specific animal models to different human OA phenotypes. DESIGN: Male 11-week-old C57BL6 mice had unilateral medial-meniscal-destabilization (DMM) or antigen-induced-arthritis (AIA). Joint tissue histopathology was scored day-3 to week-16. Tissue-pathology associations (corrected for time and at week-16) were determined by partial correlation coefficients, and odds ratios (OR) calculated for likelihood of cartilage damage and joint inflammation by ordinal-logistic-regression. RESULTS: Despite distinct temporal patterns of progression, by week-16 joint-wide OA pathology in DMM and AIA was equivalent. Significant pathology associations common to both models included: osteophyte size and maturity (r > 0.4); subchondral bone (SCB) sclerosis and osteophyte maturity (r > 0.25); cartilage erosion and chondrocyte hypertrophy/apoptosis (r > 0.4), SCB sclerosis (r > 0.26), osteophyte size (r > 0.3), and maturity (r > 0.32). DMM-specific associations were between cartilage proteoglycan loss and structural damage (r = 0.56), osteophyte maturity (r = 0.49), size (r = 0.45), and SCB sclerosis (r = 0.28). AIA-specific associations were between SCB sclerosis and chondrocyte hypertrophy/apoptosis (r = 0.40) and osteophyte size (r = 0.37); and synovitis with cartilage structural damage (r = 0.18). No tissue-pathology associations were common to both models at week-16. Increased likelihood of cartilage structural damage was associated with: chondrocyte hypertrophy/apoptosis (OR>1.7), and osteophyte size (OR>2.3) in both models; SCB sclerosis (OR = 2.0) and proteoglycan loss (OR = 2.4) in DMM; and synovitis (OR = 1.2) in AIA. Joint inflammation was associated positively with cartilage proteoglycan loss (OR = 1.4) and inversely with osteophyte size (OR = 0.21) in AIA only. CONCLUSION: This study highlights the importance of defining OA-models by initiating mechanisms and progression, not just end-stage joint-tissue pathology.

Disruption of glucocorticoid signalling in osteoblasts attenuates age-related surgically induced osteoarthritis.

OBJECTIVE: Aging is a major risk factor for osteoarthritis (OA). Skeletal expression and activity of the glucocorticoid-activating enzyme 11ß-hydroxysteroid-dehydrogenase type 1 increases progressively with age in humans and rodents. Here we investigated the role of endogenous osteocytic and osteoblastic glucocorticoid (GC) signalling in the development of osteoarthritic bone and cartilage damage in mice. METHODS: We utilized transgenic (tg) mice in which glucocorticoid signalling is disrupted in osteoblasts and osteocytes via overexpression of the glucocorticoid-inactivating enzyme, 11ß-hydroxysteroid-dehydrogenase type 2. Osteoarthritis was induced in 10- and 22-week-old male transgenic mice (tg-OA, n = 6/group) and their wildtype littermates (WT-OA, n = 7-8/group) by surgical destabilization of the medial meniscus (DMM). Sham-operated mice served as controls (WT- & tg-Sham, n = 3-5 and 6-8/group at 10- and 22-weeks of age, respectively). RESULTS: Sixteen weeks after DMM surgery, mice developed features of cartilage degradation, subchondral bone sclerosis and osteophyte formation. These changes did not differ between WT and tg mice when OA was induced at 10-weeks of age. However, when OA was induced at 22-weeks of age, cartilage erosion was significantly attenuated in tg-OA mice compared to WT-OA littermates. Similarly, subchondral bone volume (-5.2%, 95% confidence intervals (CI) -9.1 to -1.2%, P = 0.014) and osteophyte size (-4.0 mm2, 95% CI -7.5 to -0.5 mm2, P = 0.029) were significantly reduced in tg-OA compared to WT-OA mice. CONCLUSION: Glucocorticoid signalling in cells of the osteoblast lineage promotes the development of surgically-induced osteoarthritis in older, but not younger, male mice. These data implicate osteoblasts and osteocytes in the progression of DMM-OA, via a glucocorticoid-dependent and age-related pathway.

Considerations for the design and execution of protocols for animal research and treatment to improve reproducibility and standardization: "DEPART well-prepared and ARRIVE safely".

OBJECTIVE: To review the factors in experimental design that contribute to poor translation of pre-clinical research to therapies for patients with osteoarthritis (OA) and how this might be improved. METHODS: Narrative review of the literature, and evaluation of the different stages of design conduct and analysis of studies using animal models of OA to define specific issues that might reduce quality of evidence and how this can be minimised. RESULTS: Preventing bias and improving experimental rigour and reporting are important modifiable factors to improve translation from pre-clinical animal models to successful clinical trials of therapeutic agents. Despite publication and adoption by many journals of guidelines such as Animals in Research: Reporting In Vivo Experiments (ARRIVE), experimental animal studies published in leading rheumatology journals are still deficient in their reporting. In part, this may be caused by researchers first consulting these guidelines after the completion of experiments, at the time of publication. This review discusses factors that can (1) bias the outcome of experimental studies using animal models of osteoarthritis or (2) alter the quality of evidence for translation. We propose a checklist to consult prior to starting experiments; in the Design and Execution of Protocols for Animal Research and Treatment (DEPART). CONCLUSIONS: Following DEPART during the design phase will enable completion of the ARRIVE checklist at the time of publication, and thus improve the quality of evidence for inclusion of experimental animal research in meta-analyses and systematic reviews: "DEPART well-prepared and ARRIVE safely".

Utility of circulating serum miRNAs as biomarkers of early cartilage degeneration in animal models of post-traumatic osteoarthritis and inflammatory arthritis.

OBJECTIVE: The purpose of this study was to determine if serum microRNA (miRNA) signatures were biomarkers of early cartilage degeneration in preclinical mouse models of post-traumatic osteoarthritis (OA) and inflammatory arthritis. METHODS: Cartilage degeneration was induced in 10-12 week old male C57BL6 mice by destabilization of the medial meniscus (DMM) or intra-articular injection of methylated-bovine-serum-albumin (AIA), with sham-operated or saline-injected control animals (n = 6/treatment/time). Total serum RNA and knee joints were isolated at 1, 4 and 16 weeks post-induction. Cartilage degeneration was scored histologically. Serum miRNA expression profiling was performed using Agilent microarrays and validated by qPCR. RESULTS: DMM-operated and AIA mice had characteristic cartilage degeneration (proteoglycan loss, chondrocyte hypertrophy, structural damage), that increased significantly with time compared with controls, and with distinct temporal differences between arthritis models. However, expression profiling revealed no statistically significant dysregulation of serum miRNAs between AIA vs saline-injected or DMM vs sham-operated control mice at the critical early disease stages. The inability to detect DMM or AIA serum miRNA signatures compared with controls was not due to the insensitivity of the expression profiling approach since significant changes were observed in miRNA expression between the arthritis models and between time points. CONCLUSION: While distinct patterns of progressive cartilage degradation were induced in the arthritis models, we were unable to identify any serum miRNAs that were significantly dysregulated in early stages of disease compared with controls. This suggests circulating serum miRNAs may not be useful as cartilage biomarkers in distinguishing the early or progressive stages of arthritis cartilage degeneration.

Reoxygenation enhances tumour necrosis factor alpha-induced degradation of the extracellular matrix produced by chondrogenic cells.

Mesenchymal stem cells (MSCs) have been considered as a potential source for cell-based therapies in arthritic diseases for both their chondrogenic and anti-inflammatory properties. Thus, we examined how MSC-based neocartilage responds to tumour necrosis factor alpha (TNF-α) compared to articular chondrocyte (AC)-based neocartilage. Since oxygen tension is altered in arthritic joints, we also examined how increased oxygen tension influences this process. Monolayer-expanded healthy human ACs and bone marrow MSCs were cultured in chondrogenic medium in three-dimensional culture under hypoxia. They were then exposed to TNF-α under hypoxic or increased oxygen tension. We found no inherent anti-inflammatory potential of MSC-derived neocartilage as it pertains to the enzymes studied here: more degradative enzymes were upregulated by TNF-α in MSCs than in ACs, regardless of the oxygen tension. MSCs were also more sensitive to reoxygenation during TNF-α exposure, as indicated by increased proteoglycan loss, increased aggrecanase-generated metabolites, and further upregulation of the major aggrecanases, ADAMTS4 and ADAMTS5. There was also evidence of matrix metalloproteinase (MMP)-mediated aggrecan interglobular domain cleavage and type II collagen loss in response to TNF-α in both MSCs and ACs, but more MMPs were further upregulated by reoxygenation in MSCs than in ACs. Our study provides further evidence that consideration of oxygen tension is essential for studying cartilage degradation; for example, neocartilage produced from MSCs may be more sensitive to the negative effects of repeated hypoxia/reoxygenation events than AC-derived neocartilage. Consideration of the differences in responses may be important for cell-based therapies and selection of adjunctive chondroprotective agents.

Pericellular colocalisation and interactive properties of type VI collagen and perlecan in the intervertebral disc.

The aim of this study was to immunolocalise type VI collagen and perlecan and determine their interactive properties in the intervertebral disc (IVD). Confocal laser scanning microscopy co-localised perlecan with type VI collagen as pericellular components of IVD cells and translamellar cross-bridges in ovine and murine IVDs. These cross-bridges were significantly less abundant in the heparin sulphate deficient Hspg2 exon 3 null mouse IVD than in wild type. This association of type VI collagen with elastic components provides clues as to its roles in conveying elastic recoil properties to annular tissues. Perlecan and type VI collagen were highly interactive in plasmon resonance studies. Pericellular colocalisation of perlecan and type VI collagen provides matrix stabilisation and cell-matrix communication which allows IVD cells to perceive and respond to perturbations in their biomechanical microenvironment. Perlecan, at the cell surface, provides an adhesive interface between the cell and its surrounding extracellular matrix. Elastic microfibrillar structures regulate tensional connective tissue development and function. The 2010 Global Burden of Disease study examined 291 disorders and identified disc degeneration and associated low back pain as the leading global musculoskeletal disorder emphasising its massive socioeconomic impact and the need for more effective treatment strategies. A greater understanding of how the IVD achieves its unique biomechanical functional properties is of great importance in the development of such therapeutic measures.

A commentary on modelling osteoarthritis pain in small animals.

OBJECTIVE: To describe the currently used animal models for the study of osteoarthritis (OA) pain, with an emphasis on small animals (predominantly mice and rats). OUTLINE: Narrative review summarizing the opportunities and limitations of the most commonly used small animal models for the study of pain and pain pathways associated with OA, and discussing currently used methods for pain assessment. Involvement of neural degeneration in OA is briefly discussed. A list of considerations when studying pain-related behaviours and pathways in animal models of OA is proposed. CONCLUSIONS: Animal models offer great potential to unravel the complex pathophysiology of OA pain, its molecular and temporal regulation. They constitute a critical pathway for developing and testing disease-specific symptom-modifying therapeutic interventions. However, a number of issues remain to be resolved in order to standardize pre-clinical OA pain research and to optimize translation to clinical trials and patient therapies.

Comparison of gait and pathology outcomes of three meniscal procedures for induction of knee osteoarthritis in sheep.

OBJECTIVE(S): Meniscectomy (MX) of sheep induces a well-established animal model of human osteoarthritis (OA). This study compared the clinical (lameness) and pathological outcomes of unilateral, complete medial MX vs two less traumatic and more easily performed meniscal destabilisation procedures. METHODS: Four-year old wethers (n = 6/group) underwent sham operation, cranial pole release (CPR), mid-body transection (MBT) or total MX of the medial meniscus. Joints were assessed for gross pathology (cartilage erosion and osteophytes), histomorphometry, two histopathology scoring methods (modified Mankin-type and Pritzker score), and immunohistology for ADAMTS- and MMP-cleaved neoepitopes, at 12 weeks post-op. Ground reaction forces (GRFs) were determined by force plate in a subset (n = 4/group) at baseline, 2.5, 8, and 12 weeks post-op. RESULTS: Gross pathology scores of operated groups differed significantly from sham animals (P < 0.05) but not from each other, though qualitative differences were noted: CPR sheep developed more cranial and focal lesions, while MBT and MX joints showed more widespread lesions and osteophyte formation. Similarly, histopathology scores were significantly elevated vs sham but did not differ between operated groups at P < 0.05, except for a trend for lower tibial cartilage histopathology in MBT consistent with the immunohistologic pattern of reduced aggrecanase-cleavage neoepitope in that model. CPR sheep developed less femoral subchondral sclerosis, suggesting some residual biomechanical effect from the destabilised but intact meniscus. Few significant differences were noted between operated groups in force plate analyses, though gait abnormalities appeared to be least in CPR sheep, and most persistent (>12 weeks) in MBT animals. CONCLUSION: The well-validated ovine MX model and the simpler meniscal destabilisation procedures resulted in broadly similar joint pathology and lameness. Meniscal CPR or MBT, as easier and more clinically relevant procedures, may represent preferred models for the induction of OA and evaluation of potential disease-modifying therapies.

What constitutes an "animal model of osteoarthritis"--the need for consensus?

OBJECTIVE: To review the use of animal models of osteoarthritis (OA) with regard to their utility for investigation of the mechanisms and regulation of structural pathology and pain. METHODS: PubMed searches were conducted using separate clusters of terms to retrieve articles on (i) models of structural joint damage in genetically-modified (GM) mice, and (ii) models of OA joint pain. The papers were reviewed to investigate whether there was evidence that the research outcome was dependent on the model used. RESULTS: Out of a total of 109 separate GM mice strains identified in which an effect on OA was reported, 15 had been studied using more than one arthritis model. In 10/15 the same effect of the GM on arthritis was reported in at least two different models. In 5/15 the effect of the GM on arthritis structural pathology was different, and sometimes opposite, when comparing two or more induction methods. A total of 112 publications were retrieved in which pain/disability was examined in a model suggested to represent OA. The induction methods used most commonly to study "OA pain" were distinct from those most often used to investigate the pathophysiology and regulation of structural joint damage. Four papers directly comparing pain mechanisms in different models were identified, with 3/4 describing differences in nociceptive pathways. CONCLUSIONS: The available data indicates that the molecular mechanisms of both joint structural damage and pain may be distinct in animal models of OA induced or initiated by different means. This suggests the need to continue using multiple OA animal models but that the subsequent interpretation of the data and its extrapolation to the human condition must be more precise.

Zonal differences in meniscus matrix turnover and cytokine response.

OBJECTIVE: To determine the mechanisms of meniscal degeneration and whether this varied zonally and from articular cartilage. DESIGN: Normal ovine menisci were dissected into inner and outer zones and along with cartilage cultured ±IL-1α and TNFα. Glycosaminoglycan (GAG) and collagen release, and gene expression were quantified. Aggrecan proteolysis was analysed by Western blotting with neoepitope-specific antibodies. Matrix metalloproteinase (MMP)2, MMP9 and MMP13 activity was evaluated by gelatin zymography or fluorogenic assay. RESULTS: Inner meniscus was more cartilaginous containing more GAG and expressing more ACAN and COL2A1 than outer zones. Higher expression of VCAN and ADAMTS4 in medial outer and both zones of the lateral meniscus reflected their embryologic origin from cells outside the cartilage anlagen. All meniscal regions released a greater % GAG in response to cytokines; only outer zones had cytokine-stimulated collagenolysis. Cytokine-induced aggrecanolysis was primarily due to increased ADAMTS cleavage in cartilage and inner menisci but MMPs in the outer menisci. Outer menisci always released more active MMP2 than other tissues and more active MMP13 in basal and TNF-stimulated cultures. Expression of ACAN, COL1A1 and COL2A1 was decreased by both cytokines in all tissues, while VCAN was increased by IL-1α in cartilage and inner menisci. Metalloproteinase expression was differentially regulated by IL-1α and TNFα: ADAMTS4, MMP1, MMP3 were upregulated more by IL-1α in inner zones whereas ADAMTS5, MMP13 and MMP9 were more upregulated by TNFα in outer zones. CONCLUSIONS: Meniscal degeneration mechanisms are zonally-dependent, and may contribute to the enzymatic burden in the joint.

Regulation of aggrecanases from the ADAMTS family and aggrecan neoepitope formation during in vitro chondrogenesis of human mesenchymal stem cells.

Aggrecanases from the ADAMTS (A Disintegrin And Metalloproteinase with ThromboSpondin motifs) family are important therapeutic targets due to their essential role in aggrecan depletion in arthritic diseases. Whether their function is also important for matrix rearrangements during chondrogenesis and thus, cartilage regeneration, is however so far unknown. The aim of this study was to analyse the expression and function of ADAMTS with aggrecanase activity during chondrogenic differentiation of human mesenchymal stem cells (MSCs). Chondrogenic differentiation was induced in bone marrow-derived MSC pellets and expression of COL2A1, aggrecan, ADAMTS1, 4, 5, 9, 16 and furin was followed by quantitative RT-PCR. Formation of the NITEGE (ADAMTS-cleaved) and DIPEN (MMP-cleaved) aggrecan neoepitopes was detected by immunohistochemistry. While the expression of ADAMTS4, 9, 16 and furin was up-regulated during chondrogenesis, ADAMTS1 and 5 were down-regulated. Despite this regulation of ADAMTS, no formation of NITEGE neoepitopes occurred in MSC pellets, indicating no ADAMTS-induced cleavage of aggrecan. In contrast, MMP-induced cleavage of aggrecan appeared at 14 d after induction of chondrogenesis. Submission of differentiated MSC pellets to IL1ß treatment for 3 d resulted in strong upregulation of ADAMTS1, 4 and 5, rapid proteoglycan depletion, and stimulation of ADAMTS-induced but not MMP-induced cleavage of aggrecan. Thus, there is no evidence for ADAMTS-induced aggrecan cleavage during chondrogenesis, but proteoglycan turnover is rapidly inducible under inflammatory signals. Therapeutic aggrecanase inhibition for treatment of arthritic disease may thus not impede regenerative self-healing pathways based on chondrogenesis of local progenitor cells in the joint.