BACKGROUND

Although osteoarthritis (OA) is not an inflammatory arthritis, a characteristic feature of OA is increased production of pro-inflammatory cytokines, such as interleukin 1beta (IL-1b), by articular chondrocytes. In fact, the degree of articular inflammation is often associated with disease progression; indicating that this process probably contributes to articular damage. Suppressor of cytokine signalling (SOCS) proteins are, as the name suggests, inhibitors of cytokine signalling that function via the JAK/STAT pathway (Janus kinase/signal transducers and activators of transcription). Eight SOCS proteins, SOCS1-SOCS7 and CIS-1 (cytokine-inducible SH2-domain-1 with similar structure to the other SOCS proteins) have been identified, of which, SOCS1-3 and CIS-1 are the best characterised. Reduced expression of SOCS proteins would be predicted to result in increased cytokine responsiveness and thereby could contribute to OA pathology.

INTRODUCTION

Many patients suffering from osteoarthritis (OA) take daily glucosamine (GlcN) in the hope of slowing down disease progression and ameliorating pain. However, the physiological basis of this effect is not known. We previously presented preliminary data suggesting that GlcN prevented the increase in interleukin-1beta (IL-1b) expression caused by addition of inflammatory cytokines to cultures of healthy human articular chondrocytes. Previous studies had also shown that, in OA, epigenetic DNA methylation loss at specific CpG sites in relevant promoters ‘unsilences’ the genes and that this DNA de-methylation underlies the aberrant gene expression of proteases (Arthritis Rheum 52;3110-24). Furthermore, exogenous inflammatory cytokines have the capacity to cause DNA de-methylation in the IL-1b promoter (Arthritis Rheum. 2009, 60, 3303-3313). The aims of the present study were to investigate whether GlcN not only prevents the increased IL-1b expression, but also inhibits epigenetic unsilencing by preventing the cytokine-induced loss of DNA methylation.

Osteoarthritis (OA) is characterised by the progressive destruction of articular cartilage by matrix-degrading enzymes. Although initially produced by synovial fibroblasts, these enzymes are derived from OA chondrocytes as the disease progresses. Inflammatory cytokines (IL-1beta, TNF-alpha, oncostatin M) are known to induce the aberrant expression of the proteases and IL-1beta in vitro. We wanted to investigate whether the transcription factor NF-kB, which is frequently involved in signal transduction, is a mediator of the effects of inflammatory cytokines. Hence we determined the effects of NF-kB inhibition on the expression of IL-1beta, MMP-13 and MMP-3, which was induced in healthy chondrocytes by culturing with TNF-alpha/OsM.

Chondrocytes were isolated from the healthy cartilage of femoral heads obtained from patients after hemiarthroplasty following a femoral neck fracture (n=4). The chondrocytes from each patient were divided into four experimental groups: untreated control culture; culture with TNF-alpha/OsM; culture with TNF-alpha/OsM in presence of an NF-kB inhibitor; and culture with TNF-alpha/OsM treated + control peptide. Two inhibitors of NF-kB nuclear translocation were employed: an NF-kB p65 (ser276) inhibitory peptide (Imgenex) and (E)-2-Fluoro-4′-methoxystilbene, an analogue of resveratol (Merck). Cells were grown in monolayer culture for two weeks and received two rounds of treatment. Once confluent, cells were harvested and total RNA was extracted, using a Qiagen kit. RNA was reverse transcribed into cDNA and the expression of IL-1beta, MMP-3 and MMP-13 was analysed by conventional RT-PCR.

No expression of IL-1beta was found in control cultures but expression was induced, as expected, following treatment with TNF-alpha/OsM. Presence of the NF-kB inhibitor reduced IL-1beta expression, but did not abolish it completely, as suggested by reduced intensity of the PCR band. This was seen in all four samples. Similarly, NF-kB inhibition attenuated MMP-13 expression in three patients, but in one patient MMP-13 was already expressed in control cultures and no change was observed in the treated groups. MMP-3 was uniformly expressed across all experimental groups and was unaffected by NF-kB inhibition.

NF-kB is the generic name for a family of transcription factors, of which the p65-p50 heterodimer is the most prevalent. NF-kB is normally sequestered within the cytoplasm in an inactive form by binding to inhibitory kB (IkB) proteins. Activation involves degradation of IkB and nuclear translocation of NF-kB. The present results show that the cytokine-induced expression of IL-1beta and MMP-13 in healthy chondrocytes involves nuclear translocation of NF-kB.

Osteoarthritis (OA) is a common degenerative disease associated with aging thatas yet has no cure. Glucosamine (Gln) is a naturally produced amino sugar that forms part of the cartilage matrix and is taken by millions of OA sufferers in the hope of alleviating their symptoms. Apart from alleviating pain, there is evidence in the literature that Gln may also be a chondroprotective drug in OA and some clinical trials have shown reduced joint space narrowing in patients taking 1mg Gln per day. However, the mechanisms by which Gln might have its beneficial effects are still uncertain.

We wanted to determine whether Gln has any influence on the aberrant gene expression that takes place in OA chondrocytes. To this end, we cultured healthy articular chondrocytes and induced aberrant gene expression with TNF-α /OSM. Healthy human chondrocytes were isolated from the cartilage of the femoral head obtained after hemiarthroplasty from four patients who had fractured the neck of their femur. Each sample was divided in to 4 groups prior the monolayer culture:

Control culture,

At confluency (~ 2 weeks) RNA was extracted for analysis of mRNA expression by RT-PCR. The impact of Gln on the expression if the inflammatory cytokine IL-1b and the protease MMP-13 was determined by conventional RT-PCR.

No expression of IL-1b was found in control cultures and Gln on its own did not induce expression. As expected, TNF-a/OSM induced the expression of IL-1b in all four patients. When Gln was present together with TNF-a/OSM, IL-1b expression was prevented in two patients and considerably reduced in the other two patients. With respect to MMP-13, expression was present in 3/4 cultured controls and Gln did not influence this expression. TNF-α /OSM increased expression of IL-1b, and the cytokine-induced expression was slightly reduced by Gln in 2/4 patients.

These results suggest that Gln prevents the TNF-α /OSM-induced expression of IL-1b, but has limited direct influence on MMP-13 expression, at least in vitro. If the data are applicable to the in vivo situations, the results support the proposed chondroprotective effect of glucosamine at the cellular level.

Epigenetic DNA de-methylation at specific CpG promoter sites is associated with abnormal synthesis of matrix-degrading enzymes in human osteoarthritis (Arthritis Rheum 52:3110–24), but the mechanisms that trigger or cause loss of DNA methylation are not known. Since inflammatory cytokines are known to induce abnormal gene expression in cultured chondrocytes, we wanted to know whether this induction also involved loss of DNA methylation. If so, the abnormal gene expression would be permanent and transmitted to daughter cells rather than a simple up-regulation. To test this hypothesis, we selected IL-1b as the abnormally expressed gene. Healthy chondrocytes, harvested from human femoral head cartilage following a fracture, were divided into five groups: non-culture; control culture; culture with the de-methylating agent 5-aza-deoxycyti-dine (5-aza-dC); culture with the inflammatory cytokine IL-1b; or with TNF-a/oncostatin M. Total RNA and genomic DNA were extracted at confluency, relative mRNA expression of IL-1b was quantified by Syb-rGreen-based real-time PCR, and a method for quantifying the percent of cells with DNA methylation at a specific CpG site was developed (Epigenetics 2: 86–95).

The methylation status of 16 CpG sites in the promoter of IL-1b was determined by the bisulfite modification method. The two CpG sites important for the epigenetic regulation of IL-1b were at -247bp and -290bp, the latter was selected to quantify DNA methylation. 5-aza-dC halved DNA methylation, which resulted in 4–8 fold increases in IL-1b expression; showing that DNA de-methylation per se increases gene expression. However, far greater effects were seen with the inflammatory cytokines. IL-1b increased its own expression 50–100 fold, whereas TNF-a/OSM increased IL-1b expression 500–1000 fold. DNA methylation varied inversely, IL-1b reducing methylation to ~15% and TNF-a/OSM abolishing DNA methylation almost completely.

This is the first demonstration that inflammatory cytokines have the capacity to cause loss of DNA methylation. We also confirmed previous work that IL-1b induces its own expression in healthy chondrocytes, thus setting up a dangerous positive feed-back mechanism. If true in vivo, both the auto-induction and the heritable expression of IL-1b by a growing number of chondrocytes could explain the unrelenting progression of osteoarthritis.

Regenerative medicine provides the hope for many intractable diseases as a treatment option and the area is currently the subject of intense investigation in academia and industry. Human bone marrow stromal cells (HBMSCs) possess the ability to differentiate into a variety of cell types of the stromal lineage including cells of the osteogenic and chondrogenic lineages. However, the process of in vitro differentiation is usually inefficient, difficult to reproduce in many cases and, to date, unable to produce homogenous cell populations, which is critical for tissue engineering. Epigenetic regulation of gene expression is recognized as a key mechanism governing cell determination, commitment, and differentiation as well as maintenance of those states. The main components of epigenetic control are DNA methylation and histone acetylation. During development, the epigenetic status changes as cells differentiate along specific lineages. We reasoned that epigenetic modifiers might direct the differentiation pathway of HBMSCs towards either osteogenic or chondrogenic lineage. HBMSCs were serum-starved for 24 hours to synchronise the cell cycle, then treated on three consecutive days either with the DNA demethylating agent 5-Aza-deoxycytidine (5-Aza-dC) 1?M, or the histone deacetylase inhibitor Trichostatin A (TSA) 100 nM or a combination of both. After confluency, the cells were grown in pellet culture for 21 days to facilitate formation of an extracellular matrix. 5-Aza-dC increased the amount of osteoid in the pellet by at least 5 fold compared with controls as assessed by histochemistry, whereas TSA enhanced formation of a cartilage matrix. The differentiation was further enhanced by culturing the pellets in osteogenic or chondrogenic media. These studies suggest that loss of DNA methylation stimulates osteogenic differentiation, whereas inhibition of histone deacetylation favours chondrogenesis. Epigenetic changes thus play an important role in HBMSCs differentiation and offer new approaches in skeletal tissue engineering programs. The challenge will be to define the crucial genes in which loss of DNA methylation has taken place or how changes in histone acetylation (and other histone modifications) affect lineage differentiation.

Osteoarthrosis (OA) is often considered to be due to “wear and tear”, aggravated by obesity. However, if developmental dysplasia of the hip (DDH) is treated incorrectly, osteoarthrosis can also occur at a very young age. We obtained cartilage from the femoral head a 23 year-old female after arthroplasty for DDH; from a 14 year-old male, resected for paralytic dislocation, and from OA and control patients. This provided a unique opportunity to compare the cellular and epigenetic features of OA in older patients with those in a young control as well as a DDH patient. We have recently defined the cellular and epigenetic features of idiopathic OA, in particular the association between induction of proteases and loss of DNA methylation in the respective promoter regions (Arthritis Rheum2005; 52: 3110–3124). We had shown that these proteases were silenced in normal articular chondrocytes, but “unsilenced” in OA chondrocytes. The present study determined whether the phenotypic changes of idiopathic OA also take place in juvenile OA and whether loss of DNA demethylation is also associated with the abnormal expression of proteases in juvenile OA. Paraffin sections were immunostained with antibodies to MMP-3, -9, and ADAMTS-4. DNA was extracted from freezer milled cartilage. The methylation status of specific CpG sites (at which methylation occurs) was established using methylation-sensitive restriction enzymes followed by PCR. From the 23 year old female, we only obtained a 1cm thick transverse slice of femur, taken near the femoral neck. However, this contained sufficient reasonably thick cartilage in the perimeter for histology and DNA extraction.

The cartilage of the 14-year old showed high cellularity and absence of immunostaining for all proteases investigated. Apart from the higher cellularity, this was similar to the ‘control’ cartilage obtained from patients with a fracture of the femoral neck. We had previously shown that, as OA progresses, more chondrocytes become immunopositive for the degradative enzymes and these cells divide so that in the typical clones of OA all cells synthesize the proteases. The cartilage from the 23-year old DDH patient showed extensive loss of proteoglycans from the superficial zone and fibrous repair tissue covered some areas. Nearly all chondrocytes produced the proteases and clones had formed, as in idiopathic OA. Since these sample were from the base of the femoral neck, where in idiopathic OA good cartilage often remains even in severe OA, the disease process must have reached an early end-stage in this young patient. The findings indicate that severe OA, as defined by the presence of clones that produce degradative enzymes, can develop very quickly. Interestingly, the expression and synthesis of degradative enzymes by OA chondrocytes was the same in juvenile and old-age OA. and their abnormal expression was associated with “unsilencing” via DNA demethylation in both juvenile and old-age OA. The results thus suggest that age per se is not a major determinant of OA progression.

Cartilage and bone degeneration are major healthcare problems affecting millions of individuals worldwide. Elucidation of the processes modulating the cell-matrix interactions involved in cartilage or bone formation offer tremendous potential in the development of clinically relevant strategies for cartilage and bone regeneration. We have therefore adopted an ex vivo tissue engineering approach to investigate chondrogenesis and osteogenesis using a mix human mesenchymal progenitor populations encapsulated in biomineralised polysac-charide templates with or without the addition of type-I collagen.

Alginate/chitosan polysaccharide capsules containing 2.5mg/ml type-I collagen and TGF-beta-3 were encapsulated with human bone marrow cells (HBMC), articular chondrocytes or a co-culture at a ratio of 2:1 respectively and placed in a rotating (Synthecon) biore-actor or held in static 2D culture conditions for 28 days, to determine whether the presence of type-I collagen within the alginate could promote the synthesis of an extracellular matrix.

Constructs were stained with alcian blue, sirius red and von Kossa. In bioreactor samples encapsulated with HBMC and type-I collagen, viable cells were present within lacunae, surrounded by a matrix of proteo-glycans and fibrous collagen, which was mineralized. Immunohistochemistry and polarised light microscopy indicated an organised collagenous matrix with extensive expression of type I collagen and bone sialoprotein with small regions of type II collagen. Type X collagen was also expressed indicating the presence of hypertrophic chondrocytes. Within the static HBMC groups, smaller areas of matrix were generated with decreased expression of type-I and type-II collagen. Co-culture bioreactor samples also demonstrated regions of new mineralised bone matrix; however these were less prominent than in the HBMC only groups. No matrix formation was observed in chondrocyte cultures although the cells remained viable as assessed by live/dead staining. Biochemical analysis indicated significantly increased (p< 0.05) DNA in all bioreactor samples in comparison with static constructs and significantly increased protein in HBMC bioreactor constructs in comparison with other cell types.

These studies outline a unique tissue engineering approach, utilizing individual and mixed human mesen-chymal progenitor populations coupled with innovative polysaccharide templates containing type I collagen and bioreactor systems to promote chondrogenic and osteo-genic differentiation.

Osteoarthritis (OA) is characterised by the progressive loss of the articular cartilage. This is accompanied by change in phenotype from cells expressing chondrocytic genes to cells, termed ‘degradative’ chondrocytes, that express aggrecanases and collagenases. To understand the cellular events involved, human articular cartilage was obtained from femoral heads after arthroplasty due to OA, fracture of the neck of femur (#NOF) due to osteoporosis, or from a 14 year old male (CDH). Samples were graded according to the new OARSI system (Osteoarthritis and Cartilage, 2006, 14:13–29) and par-affin sections were immunostained for c-Myc (marker of cellular activation), S100 (typically expressed in chon-drocytes), Sox-9 (expressed in early-stage chondrocytes) and nucleostemin (a stem-cell marker). In addition, some specimen were incubated with fluorescent probes to identify metabolically activated cells (CellTracker green).

All chondrocytes, irrespective of OA grade, were immunopositive for S100, but there were differences in the other parameters. Cartilage from the 14-year old (OARSI grade =0) was characterized by no loss of proteoglycans (safranin-O) in the superficial zone and absence of c-Myc, Sox-9 and nucleostemin in all articular chondrocytes. In #NOF cartilage, proteoglycan loss was evident in the very superficial zone. Many chondro-cytes in that zone showed bright green fluorescence with CellTracker-green and were c-Myc positive, consistent with cellular activation. Sox-9 and nucleostemin were absent. Mid-zone and deep zone chondrocytes showed no change. In low-grade OA samples (OARSI = 1-2), the zone of proteoglycan loss had increased, the Cell-Tracker-green/c-Myc positive chondrocytes in that zone had divided to form clusters of 4-8 cells. Occasional cells were positive for nucleostemin. Mid-zone and deep zone chondrocytes still showed no change. In high-grade fib-rillated OA cartilage (OARSI = 3-4) the superficial and mid zones had been eroded, leaving the deep zone at the surface. Chondrocytes were typically found in large clones, which were all immunopositive for c-Myc as well as for nucleostemin and Sox-9.

The results show that cellular activation starts near the surface and progresses to the deep zone. The presence of nucleostemin and Sox-9 suggests that de-differentiation may be involved in the phenotypic change from the chondrocytic to the degradative phenotype.

Osteoarthritis (OA) is characterized by progressive erosion of articular cartilage due to degradation of the cartilage matrix. The major enzymes involved are the matrix metalloproteases and aggrecanases, which are either derived from the synovium or synthesized by chondrocytes as OA progresses. This abnormal enzyme synthesis is part of a phenotypic change from normal to ‘degradative’ chondrocytes. If this change could be prevented, then disease progression might be slowed.

In early OA, degradative chondrocytes are only present in the superficial zone, but with increasing severity of OA, more chondrocytes become degradative cells so that, in high-grade OA, these cells are also located in the deep zone. We hypothesized the existence of a ‘factor X’, which diffuses from the superficial to the deep zone and induces cells to change phenotype and express the pro-teases. We further hypothesize that this factor is released by degradative chondrocytes. To test the hypothesis, we co-cultured explants of human superficial-zone OA cartilage (which contains degradative cells and thus factor X) with explants of deep-zone cartilage from fracture neck of femur patients (#NOF), which contains mostly normal chondrocytes that do not express the proteases. We investigated MMP expression by real time RT-PCR and protein synthesis by immunohistochemistry.

Before culture, MMP-2, -3, -9, or -13 were expressed in the superficial-zone OA cartilage, but not in deep-zone #NOF cartilage, as expected. After 4 weeks with separate culture of superficial zones and deep zones, no MMPs was expressed in deep zone chondrocytes, suggesting that culture per se did not induce expression of these enzymes. Neither did culture abolish expression in the superficial zone, as confirmed by RT-PCR and immunohistochemistry. However, when superficial-zone cartilage was co-cultured with deep-zone cartilage, MMP-3 expression were induced in deep- zone chon-drocytes, suggesting that a diffusible factor X, derived from degradative chondrocytes, had induced normal articular chondrocytes to express MMP-3. These experiments provide evidence for the existence of a factor that, when diffusing through the cartilage matrix, has the potential to induce normal non-enzyme expressing cells to become degradative chondrocytes.

Introduction: Osteoarthritis (OA) has historically been thought of as a degenerative joint disease, but inflammation and angiogenesis are increasingly being recognised as contributing to the pathogenesis, symptoms and progression of OA. β-dystroglycan (β-DG) is a pivotal element of the transmembrane adhesion molecule involved in cell-extracellular matrix adhesion and angiogenesis. Matrix metalloproteinases (MMPs) are the main enzymes responsible for cartilage extracellular matrix breakdown and are also implicated in both angiogenesis and β-DG degradation in a number of malignancies. We aimed to investigate the expression and localisation of β-DG and MMP-3, -9, and -13 within cartilage, synovium and synovial fluid and establish their roles in the pathogenesis of OA.

Methods: Following ethical committee approval, cartilage, synovium and synovial fluid were obtained from the hip joints of 5 osteoarthritic (patients undergoing total hip replacement) and 5 control hip joints (patients undergoing hemiarthroplasty for femoral neck fracture). The samples were analysed for β-DG expression using Western Blotting and for the distribution of β-DG, MMP-3, -9, and -13 using immunohistochemistry on paraffin embedded tissue.

Results: Whilst no significant expression of β-DG was found in cartilage or synovial fluid, β-DG was expressed in the smooth muscle of both normal and osteoarthritic synovial blood vessels. Moreover, β-DG was expressed in endothelium of blood vessels of OA synovium, but not in the normal endothelium. In the endothelium of osteoarthritic synovial blood vessels, β-DG co-localised with MMP −3 and −9.

Discussion: Our results demonstrate that β-DG does not act as a cell adhesion molecule binding chondrocytes to the ECM. However, specific immunolocalisation of β-DG within endothelium of inflamed OA blood vessels suggests that β-DG may play a role in angiogenesis associated with OA. Its co-localisation with MMP-3 and −9, previously reported to also have pro-angiogenic roles, may be linked. Further research is required to understand these roles more fully.

Introduction: Osteoarthritis has historically been thought of as a degenerative joint disease, but inflam-mation and angiogenesis are increasingly being recognised as contributing to the pathogenesis, symptoms and progression of osteoarthritis. Beta-dystroglycan is a pivotal element of the transmembrane adhesion molecule involved in cell-extracellular matrix adhesion and angiogenesis. Matrix metalloproteinases are the main enzymes responsible for cartilage extracellular matrix breakdown and are also implicated in both angiogen-esis and beta-dystroglycan degradation in a number of malignancies. We aimed to investigate the expression and localisation of beta-dystroglycan and matrix metal-loproteinase 3, 9, and 13 within cartilage, synovium and synovial fluid and establish their roles in the pathogenesis of osteoarthritis.

Methods: Following ethical committee approval, cartilage, synovium and synovial fluid were obtained from the hip joints of 5 osteoarthritic (patients undergoing total hip replacement) and 5 control hip joints (patients undergoing hemiarthroplasty for femoral neck fracture). The samples were analysed for beta-dystroglycan expression using Western Blotting and for the distribution of beta-dystroglycan, matrix metalloproteinase 3, 9, and 13 using immunohistochemistry on paraffin embedded tissue.

Results: Whilst no significant expression of beta-dystro-glycan was found in cartilage or synovial fluid, beta-dys-troglycan was expressed in the smooth muscle of both normal and osteoarthritic synovial blood vessels. Moreover, beta-dystroglycan was expressed in endothelium of blood vessels of osteoarthritic synovium, but not in the normal endothelium. In the endothelium of osteo-arthritic synovial blood vessels, beta-dystroglycan co-localised with matrix metalloproteinase 3 and 9. Discussion: Our results demonstrate that beta-dystro-glycan does not act as a cell adhesion molecule binding chondrocytes to the extracellular matrix. However, specific immunolocalisation of beta-dystroglycan within endothelium of inflamed osteoarthritic blood vessels suggests that beta-dystroglycan may play a role in angiogenesis associated with osteoarthritis. Its co-localisation with matrix metalloproteinase 3 and 9, previously reported to also have pro-angiogenic roles, may be linked. Further research is required to understand these roles more fully.

Idiopathic osteoarthritis (OA) is a complex, late-onset disease whose causes are still unknown. In spite of tremendous efforts, the search for the genes pre-disposing towards osteoarthritis has so far met with little success. We hypothesize that epigenetic changes play a major role in the pathology of OA. Epigenetics refers to stable, heritable, but potentially reversible modifications of gene expression that do not involve mutations in the DNA sequence, for example DNA methylation or histone modification. Epigenetic changes are gene and cell-type specific, may arise sporadically with increasing age or be provoked by environmental factors. To investigate whether epigenetic changes are significant factors in OA, we examined the DNA methylation status of the promoter regions of three genes that are expressed by OA, but not by normal, articular chondrocytes, namely MMP-3 (stromelysin-1), MMP-9 (gelatinase B) and MMP-13 (collagenase3). We hypothesized that these genes are silenced in normal chondrocytes by methylation of the cytosines of CpG dinucleotides in the respective promoter regions, but that abnormal expression is associated with a de-methylation, leading to eunsilencing f of gene expression. Cartilage was obtained from the femoral heads of 16 OA and 10 femoral neck fracture (#NOF) patients, which served as controls due to the inverse relationship between osteoporosis and OA. The cartilage was milled in a freezer mill with liquid nitrogen, DNA was extracted with a Qiagen kit, digested with methylation sensitive restriction enzymes, followed by PCR amplification. These enzymes will cut at their specific cleavage sites only if the CpGs is not methylated and thus allow us to determine methylation status of specific CpG sites.

Results. Less than 5% of the chondrocytes in superficial layer from #NOF cartilage expressed degradative enzymes, whereas all cloned chondrocytes from advanced-stage OA cartilage were immunopositive. The overall % of CpG demethylation in the promoters of control patients (whose chondrocytes did not express the enzymes) was 20.1%, whereas 48.6% of CpG sites were demethylated in degradative chondrocytes of OA patients (p< 0.001). For MMP-13, the increase in demethylation between control and OA was from 4 ^..20%; for MMP-9 from 47 ^..81% and for MMP-3 from 30 ^..57%. However, not all available CpG sites were equally demethylated. Some sites were uniformly methylated in both OA and controls, others were demethylated even in controls. However, there was at least one crucial site for each degradative enzyme, where the differences in the degree of methylation were greatest and statistically different. These sites were at −110 for MMP-13; −36 for MMP-9; −635 for MMP-3. There was no relation between the % demethylation and the patient fs age and no apparent difference between males and females.

Conclusions: We have demonstrated an association between abnormal gene expression of MMP-3, MMP-9 and MMP-13 and promoter DNA demethylation. This epigenetic dysregulation of genes appeared to be clonally inherited by daughter cells and may be typical for osteoarthritis and other complex, late-onset diseases.

Evidence has accumulated in recent years that programmed cell death (PCD) is not necessarily synonymous with the classical apoptosis, as defined by Kerr & Wyllie (J Path, 1973, 111:255–261), but that cells use a variety of pathways to undergo cell death, which are reflected by different morphologies. Although chondrocytes with the hallmark features of classical apoptosis have been demonstrated in culture, such cells are extremely rare in vivo. We have examined the morphological differences between dying chondrocytes and classical apoptotic cells in growth plate and osteoarthritic chondrocytes. Unlike classical apoptosis, chondrocyte death involves an increase in the endoplasmic reticulum and Golgi apparatus. This is likely to reflect an increase in protein synthesis with retention of proteins in the ER leading to expansion of the ER lumen, whose membranes surround and compartmentalise organelles and parts of cytoplasm. The final removal of apoptotic remains does not involve phagocytosis, but a combination of three routes: 1) auto-digestion of cellular material within compartments formed by ER membranes; 2) autophagic vacuoles and 3) extrusion of cell remnants into the lacunae. Together these processes lead to complete self-destruction of the chondrocyte as evidenced by the presence of empty lacunae. The involvement of ER suggests that the endoplasmic reticulum pathway of apoptosis may play a greater role in chondroptosis than receptor-mediated and mitochondrial pathways. Lysosomal proteases, present in autophagic digestion, are likely to be as important as caspases in the programmed cell death of chondrocytes in vivo. We propose the term ‘chondroptosis’ to reflect the fact that such cells are undergoing apoptosis, albeit in a non-classical manner, but one that appears to be typical of programmed chondrocyte death in vivo. Chondroptosis may serve to eliminate cells that are not phagocytosed by neighboring cells, which constitutes a crucial advantage for chondrocytes that are typically embedded in an extracellular matrix. Classical apoptosis in that situation is likely to lead to secondary necrosis with all its disadvantages. This may be the reason why most programmed cell death of chondrocytes in vivo appears to follow a chondroptotic pattern and not the classical apoptotic pattern. At present the initiation factors or the molecular pathways involved in chondroptosis remain unclear.

Clonal chondrocytes of osteoarthritic (OA) cartilage express an aberrant set of genes. We hypothesize that this aberrant gene expression may be due to clonally inherited epigenetic changes, defined as altered gene expression without changes in genetic sequence. The major epigenetic changes are due to altered DNA methylations in crucial parts of the promoter region. If the cytosines of CpG dinucleotides are methylated, the gene will be silenced, even if the right transcription factors are present. Similarly, de-methylations may activate previously silenced genes. Our aims were to provide ‘proof-of-concept’ data by examining the methylation status of genes in OA vs non-OA chondrocytes. Articular cartilage was obtained a) from the cartilage of fracture-neck-of-femur (#NOF) patients and b) from or around the eroded regions of OA samples. The former was full thickness cartilage, the latter was partially degraded cartilage, which contained mostly clonal chondrocytes as confirmed by histology. The cartilage samples were ground in a freezer mill (Glen Creston, UK) and DNA was extracted with a Qiagen DNeasy maxi kit. To assess DNA methylation status, the genomic DNA was treated overnight with methylation-sensitive restriction enzymes. Cleavage of selected sites was detected by PCR amplifications with primer pairs designed to bracket selected promoter regions. Loss of the PCR band after digestion with the enzymes indicated absence of methylations, whereas presence of the band indicated methylated cytosine. We selected MMP-9 as one of genes that is activated in OA. Transcription of mmp-9 is regulated by a 670 bp sequence at the 5′-end flanking region, which contains 6 CpGs and a further 21 CpGs within the 1.5 kb region further upstream. A PCR primer pair was designed to bracket a 350bp sequence upstream from the transcription start site of mmp-9, which contained four of the six potential methylation sites, cleaved by the methylation-sensitive enzymes AciI and HhaI. DNA from 9 OA patients, 5 #NOF patients and 1 rheumatoid arthritic (RA) patient were digested with HhaI or AciI and examined for the presence or absence of PCR bands. In all patients, digestion with HhaI abolished the PCR band, indicating that the HhaI site was never methylated in either #NOF or OA patients. However, a remarkable difference was found after digestion with AciI: in 8/9 OA patients, the PCR band was no longer detectable, while in 4/5 #NOF patients the PCR band was still present. This suggested that all three AciI cleavage sites were methylated in the majority of chondrocytes from #NOF patients, while at least one of the three AciI cleavage sites was unmethylated in OA patients. Interestingly, the PCR band was present in the RA patient, suggesting methylation of the AciI cleavage sites. The present study provides the first ‘proof-of-concept’ data that suggest epigenetic changes may play a role in the etiology of osteoarthritis. Clearly further work is required to establish the generality of the present findings and whether de-methylations are also found in the promoter regions of other genes that are aberrantly expressed in OA.

Introduction: Chondro-epiphyseal cartilage is generally resistant to vascular invasion. At the time of formation of the secondary ossification center in skeletal ‘long’ bones, the anti-angiogenic nature of cartilage is altered in favor of angiogenesis and vascular invasion takes place. We studied the control of this angiogenic ‘switch’ by experimentally investigating two factors which might influence vascular invasion. MMP 9 is a 92Kda gelatinase which degrades collagen types IV, V and X and gelatin (denatured collagen). It has been implicated in the control of endochondral ossification at the growth plate and has been shown to modulate endothelial cell morphogenesis. Basic Fibro-blast Growth Factor (b-FGF) is a cytokine with well established angiogenic capability and has also been implicated in the development of the growth plate. We investigated whether MMP-9 caused an effect on the development of the vasculature of the chondro-epiphysis of neo-natal rabbits and compared this to the effects of b-FGF.

Materials and Methods: The CAM Culture consists of placing a small tissue explant onto the the chorioallantoic membrane of 10 day-old chick embryos and continuing culture for a further 10 days. CAM derived vessels will invade the tissue, unless anti-angiogenic factors are present. Hence, CAM culture is used as an assay system for angiogenesis and factors that will influence it. We utilized the CAM culture model to investigate vascular in-growth into explants of femoral and humeral heads from 4 day old postnatal rabbits to test the influence of MMP-9 and b-FGF. A small nylon membrane, pre-soaked in a solution containing the factor, was placed on to a tangential cut across the perichondrium. The explant was then cultured on the CAM for 3–10 days.

Results: In control epiphyses, the in-growth of CAM derived blood vessels was rare and invasion of cartilage canals through the perichondrium seldom occurred, thus confirming the anti-angiogenic nature of epiphyseal cartilage. The initial presence of MMP 9 caused a tremendous increase in the de novo vascular invasion. MMP 9 treated epiphyses contained numerous large cartilage canals. In b-FGF treated epiphyses, a greater level of vascular in-growth was seen compared with controls, but this was not as marked as with MMP 9.

Our findings indicate that b-FGF and perhaps, more interestingly, MMP-9 are implicated in the activation of the angiogenic ‘switch’ at the chondroepiphysis leading to vascular invasion. The fact that MMP-9 can act as a stimulator to angiogenesis is a novel finding. The mechanism of action remains unclear although it is possible that it is involved in the deactivation of inhibitors of vasculogenesis or the activation of angiogenic factors, or both.

Programmed cell death (PCD) contributes to the pathogenesis of many diseases including osteoarthritis. The principal method is apoptosis that has a well-defined and very characteristic morphology and biochemistry.

The aim of the present study was examine whether the mechanism of cell death in OA chondrocytes was classical apoptosis.

Rat thymocytes were used as controls since these cells are known to undergo classical apoptosis. Human OA cartilage was obtained from femoral head of patients (50 – 80 years) who were undergoing joint replacement surgery. Pieces of OA samples were processed into paraffin and sections incubated with the following antibodies: M3O, an antibody that recognizes the cleavage of cytokeratin 18 by caspases; annexin V, which recognizes phosphatidylserine “flip-flop” that occurs early in the apoptotic process; bcl-2, a protein whose presence protects apoptosis and c-myc, a transcription factor thought to be associated with apoptosis. To induce apoptosis, some samples were incubated with etoposide and staurosporine.

In sections of thymus we noticed the presence of numerous apoptotic bodies. The number increased when the tissue was treated with etoposide and staurosporine. Some thymocytes were immunopositive for M3O and annexin V, and the number of positive cells increased when treated for 2h with etoposide. Chondrocytes of the articular cartilage showed chromatin condensation and many vacuoles but no fragmentation into apoptotic bodies, even when treated with etoposide or staurosporine. The OA chondrocytes were immunonegative for M3O and annexin-V, even after incubation with etoposide and staurosporine. With respect to c-myc and bcl-2, both non-weight bearing and weight-bearing areas in OA sample showed more positive cells then the thymus. More chondrocytes stained for c-myc in the superficial zone of the articular cartilage in the non-weight bearing, while in the weight-bearing areas it was more in the intermediate zone. On the other hand, there were no differences in the distribution of the cells stained for bcl-2 in the articular cartilage. It is known that some events like the phosphatidylserine flip, caspase activation and apoptotic bodies fragmentation occur quickly during apoptosis, so may be difficult to detect.

The results suggest that some features of classical apoptosis, such as phosphatidylserine flip,caspase activation and apoptotic bodies formation did not take place in OA cartilage. It is known that the molecular machinery for apoptosis is not always present in tissues that are undergoing programmed cell death, which seems to be case for OA chondrocytes.

Chondromodulin-I (ChM-I) is a bifunctional autocrine regulator of cartilage, initially isolated from fetal bovine epiphyseal cartilage¹. ChM-I stimulates matrix synthesis of chondrocytes, but inhibits the growth of endothelial cells^1,2 thus ChM-I may be one of the anti-angiogenic molecules present in cartilage. In fetal bovine bone, ChM-I was expressed by all epiphyseal and growth plate chondrocytes except hypertrophic chondrocytes and was present in the matrix throughout the epiphysis and the growth plate, except the hypertrophic zone ^2,3, consistent with its proposed role as anti-angiogenic factor. To examine the possible role of chondromodulin-I in relation to angiogenesis at the vascular front, we studied the immunolocalisation in femoral growth plates from young and mature rats (2–16 weeks) as well as aged rats (62–80 weeks), using a rabbit polyclonal antibody raised against the entire region of mature human recombinant ChM-I.

In 2-week old rats, ChM-I was synthesised by all epiphyseal chondrocytes and strong immunostaining was found in the matrix. In the growth plates, ChM-I staining was present in chondrocytes and matrix of the reserve, proliferating and maturing zones with loss of staining in the hypertrophic zone. However, ChM-I was also present where cartilage canals had penetrated into the chondroepiphysis. In 4–16 week old rats, there was a progressive change in the localisation of ChM-I. Hypertrophic chondrocytes also became positive for ChM-I, while cellular staining gradually disappeared from the other zones. By 12–16 weeks, very strong immunostaining was present almost exclusively on the inner perimeter of the lacunae of hypertrophic chondrocytes. As lacunae were opened at the vascular front, ChM-I initially remained on the cartilage-side of the lacunae, and then disappeared completely. In aged rats, very little ChM-I was present in the cells and matrix of the growth plates, except where remodelling had occurred or chondrocytes had become re-activated.

The rate of longitudinal growth in rats is high between 1–5 weeks, then declines until skeletal maturity at approximately 12 weeks, after which a very slow rate of growth continues until 26 weeks. In young rats, the loss of ChM-I in the hypertrophic zone was as expected for an anti-angiogenic factor, i.e. loss was required before vascular invasion could take place. However, the same did not apply to cartilage canal formation, since there was no loss of ChM-I around cartilage canals. The change in the localisation of ChM-I in mature rats, in particular the very intense immunolocalisation around hypertrophic chondrocytes, might be related to the reduced rate of growth. It is possible that rapid vascular invasion must be slowed down in these growth plates and that ChM-I prevented vascular invasion until degraded by proteases, such as MMP-9.

The relative absence of ChM-I in the stationary growth plates of aged rats suggests that other anti-angiogenic factors prevent vascular invasion in these growth plates.

The formation of biomimetic environments using scaffolds containing cell recognition sequence and osteo-inductive factors in combination with bone cells offers tremendous potential for bone and cartilage regeneration. In tissues, collagen forms the scaffold by mediating the flux of chemical and mechanical stimuli. Recently, a synthetic 15-residue peptide P-15, related biologically to the active domain of type I collagen, has been found to promote attachment and the osteoblast phenotype of human dermal fibroblasts and periodontal ligament fibroblasts on particulate anorganic bone mineral (ABM). The aim of this study was to exam the ability of the collagen peptide, P-15, to promote human osteoprogenitor attachment, proliferation and differentiation on cell culture surfaces and 3-D scaffolds.

Selected human bone marrow cells were cultured on particulate microporous anorganic bone mineral (‘pure ‘ hydroxyapatite based on x-ray diffraction standard JCPDS9-432) phase and polygalactin vicryl mesh adsorbed with or without P-15 in basal or osteogenic conditions. Cell adhesion, spreading and patterning were examined by light and confocal microscopy following incorporation of cell tracker green and ethidium homodimer fluorescent labels. Osteoprogenitor proliferation and differentiation was assessed by DNA content and alkaline phosphatase specific activity. Growth and differentiation on 3-D ABM structures were examined by confocal and scanning electron microscopy (SEM).

P-15 promoted human osteoprogenitor cell attachment and patterning on particulate bovine anorganic bone mineral phase and polygalactin vicryl mesh over 5–24 hours compared to culture on ABM and vicryl mesh alone as observed by photomicroscopy. Increased alkaline phosphatase specific activity was enhanced following culture on P-15 adsorbed matrices as recognized by enhanced expression of alkaline phosphatase, type I collagen, osteocalcin and cfba-1. The presence of mineralised bone matrix and extensive cell ingrowth and cellular bridging between 3-D ABM matrices and polygalactin vicryl mesh adsorbed with P-15 was observed by confocal microscopy and alizarin red staining. SEM confirmed the 3-D structure of newly formed cell constructs and cellular ingrowth on and between the P-15 modified inorganic bone mineral materials. Negligible cell growth was observed on ABM alone or polygalactin vicryl mesh alone.

These observations demonstrate that the synthetic 15-residue collagen peptide, P-15, when adsorbed to ABM or polygalactin vicryl mesh, can stimulate human osteoprogenitor attachment and spreading. They also demonstrated that P-15 coupled 3-D matrices stimulate human osteoprogenitor differentiation and materialisation. The studies indicate that a synthetic analogue of collagen provides a biomimetic environment supportive for cell differentiation and tissue regeneration and indicate a potential for the use of extracellular matrix cue in the development of biomimetic environments for bone tissue engineering.

The growth plates of rapidly growing animals have been studied extensively. Nevertheless, several questions remain unanswered, partly because many events happen simultaneously, especially at the vascular front. Terminal chondrocytes are thought to undergo programmed cell death, but the fate of the cell remnants remains unclear. Are the dying cells released into the vascular space and phagocytosed by macrophages, as one would expect for apoptosis? Or are the cells eliminated prior to opening of the lacunae, leaving empty lacunae? Do all terminal chondrocytes die or do some become bone-forming cells? Rodents maintain a growth plate into old age, long after longitudinal growth has ceased. These stationary growth plates have several features not found in the growth plates of rapidly growing animals and closer study of these features may provide answers to the above questions. Femurs and tibiae from 4–16 week-old and 62–80 week-old rats were decalcified, processed into paraffin, and the morphological changes were documented.

Between 4–16 weeks, the heights of the growth plates decreased due to loss of the large hypertrophic chondrocytes, but the various zones were still present. In the aged rats, the growth plates were identifiable as a narrow cartilaginous band with some short columns of inactive cells. The vascular front was irregular, the narrow spicules of primary spongiosa were absent and the much thicker spicules, which are normally seen in secondary spongiosa, directly abutted to the cartilage. Horizontal apposition of bone matrix onto the cartilage edge was frequently present. In addition, the following features were noted. 1) Acellular areas: Nearly all growth plates contained regions of cartilage from which all cells and their lacunae had disappeared. In some cases, these acellular regions stretched from the reserve zone to the vascular front and even persisted as a relatively wide core within the spicules of spongiosa, indicating increased resistance of acellular cartilage to resorption. The absence of cells or cell debris was consistent with an autophagic mode of cell death and subsequent collapse of the lacunae. 2) Remodelling within the growth plate; in some growth plates, large regions of growth plate cartilage had been resorbed and new bone had been laid down in a pattern similar to the remodelling of cortical bone. This suggested that the normal resistance of cartilage to vascular invasion had been lost locally, but was maintained in adjacent non-remodelled regions. 3) Trans-differentiation of chondrocytes to bone-forming cells; extensive new medullary bone formation was noted in the diaphysis of approximately 30% of the aged rats, suggesting that they had received an (unknown) osteogenic stimulus. In these rats, bone matrix was identifiable inside chondrocytic lacunae, and spreading beyond the confines of the lacunae, thus directly replacing growth plate cartilage with bone matrix.

The results suggest that i) chondrocytes are capable of self-elimination, perhaps by a mechanism similar to the autophagic cell death that occurs during insect metamorphosis; ii) resorption of cartilage and vascular invasion requires the presence of the viable chondrocytes; and iii) chondrocytes have the capacity to transdifferentiate to bone-forming cells, but only do so when receiving an increased osteogenic stimulus.