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

Destruction of articular cartilage in osteoarthritis (OA) is mediated by proteases and cytokines, which are silenced by epigenetic mechanisms in normal chondrocytes, but aberrantly expressed in OA. This is associated with DNA de-methylation of specific CpGs in the promoter regions (. Arthritis Rheum. , . 2005. ; . 52. :. 3110. –24. ). A widely used in vitro model to study the transcriptional regulation in OA is treating monolayer cultures of normal articular chondrocytes with inflammatory cytokines (IL-1b, TNFa or oncostatin M (OSM)) and investigating gene expression after 8–24 hours. The cytokines up-regulate catabolic, but down-regulate chondrocytic genes. However, whether this up- or down regulation is maintained after cytokine withdrawal is rarely investigated. In OA, the expression of catabolic genes is maintained in absence of cytokines and is transmitted to daughter cells, suggesting that epigenetic changes have resulted in permanent up-regulation. We asked whether it is possible to reproduce the epigenetic changes in vitro. Hence we compared gene expression and DNA methylation status in short-term (24h) versus long-term (2–3 weeks) cultures and, in particular, investigated the effects of cytokine withdrawal on these parameters. Healthy chondrocytes, harvested from human femoral heads after hemiarthroplasty, were cultured in monolayer and passaged once (P1). For short-term culture, the P1 chondrocytes were divided into control culture or cultures with one-shot of IL-1b/OSM, harvested after 24h and 72h. For long-term culture, the cells were cultured with or without IL-1b/OSM, the latter added twice a week. Half the cells were harvested at confluence (3 weeks) and the other halves were passaged again and cultured without cytokines until confluence (2–3 weeks). RNA and genomic DNA were extracted from the same sample. IL-1b, MMP-3, MMP-13 and COL2A1 expression was quantified by real-time PCR. The percentage of cells with DNA methylation at the CpG site at −299bp of IL-1b promoter (a key CpG site) was quantified by a method we reported previously (. Epigenetics. , . 2007. ; . 2. : . 86. –95. ). As expected, expression of IL-1b MMP-3, MMP-13 had increased 100–4500-fold 24h after IL-1b/OSM treatment, but decreased considerably after cytokine withdrawal. COL2A1 expression was virtually abolished by IL-1b/OSM and not regained after 72h. The % DNA methylation did not change during the 72h. Repeated treatment with IL-1b/OSM in long-term culture also increased expression of IL-1b and the MMPs. However, this time expression was maintained or even increased after cytokine withdrawal and passaging. Expression inversely correlated with DNA methylation, which dropped from 59% to 35%. This de-methylation was preserved after passaging and cytokine withdrawal. Conclusion: The widely used short-term cytokine-treated monolayer cultures of articular chondrocytes do not approximate the in vivo situation, where long-term aberrant expression correlates with DNA de-methylation. However, long-term treatment can mimic the loss of DNA methylation, which results in increased gene expression that is maintained after cytokine withdrawal. This model will facilitate studies on the mechanisms of DNA de-methylation, which might ultimately lead to novel therapeutic approaches for the treatment of OA