OA is a leading osteoarticular disease with declining joint functions[2, 3]. Historically, the interest in disease modification has concentrated on cartilage, which unfortunately has little direct improvement on with symptom experience[1]. The manifestations in the joint are well characterized by progressive loss and calcification of articular cartilage, subchondral bone remodeling, osteophyte formation and mild to moderate inflammation of the synovial lining[3]. It's worth noting that the synovium may show significant changes, including mononuclear cells infiltration, inflammatory cytokines production and thickening of the synovial lining layer, even before visible cartilage degeneration has occurred. In addition, a high prevalence of synovial inflammation in all stages of OA has been confirmed by the combination of high-sensitivity imaging modalities and tissue examination, with increasing studies demonstrating that synovitis is correlated with joint pain, functional impairment and may even be an independent driver of radiographic OA onset and structural progression[16]. Thus, new insights into the important roles that the synovium have in disease pathogenesis may hold great promise for future therapeutic advances.
The synovial tissues lining the diarthrodial joints surfaces include various types of cells, such as synovial fibroblasts and macrophages[17]. That these cells are involved in the pathogenesis of arthritis by producing proinflammatory cytokines, tumor necrosis factor and MMPs is widely recongnized[17, 18]. However, previous research into OA synovium was mainly based on single gene expression analysis[19–21]. The genome-wide RNA expression profiling could be useful in revealing gene functions in the disease course of such a complex syndrome. Especially pathway and network analysis would be more effective to unravel the pathogenesis of OA and even provide clues for treatment.
The expression profiling analysis via microarray can provide information about the expression differences of thousands of genes in human genome. In this study, we used it to predict and find the key or potential genes for OA. Two microarray datasets were screened (GSE82107 and GSE55235) and 300 shared DEGs obtained via bioinformatics analysis. Then, we submitted the expression matrix of all genes and phenotypes of the samples in GSE82107 to the GSEA software to perform the GO and KEGG enrichment analysis and to determine whether the DEGs show statistically significant difference between the OA and normal tissues. The GO enrichment analysis revealed that DEGs were significantly involved in cellular response to hydrogen peroxide, RNA splicing, golgi vesicle transport, cGMP binding, and lipoprotein. The KEGG enrichment analysis indicated that DEGs were generally associated with the pathway of P53, glycan, sphingolipid, ascorbate, glycerolipid, and xenobiotics by cytochrome P450.
The results obtained indicate that the cellular response to hydrogen peroxide (H2O2) of synovium membrane is involved in the disease process of OA. Besides, a previous study has suggested that H2O2 could modify the metabolism of synovial fibroblasts derived from both rheumatoid and osteoarthritic knee joints, and that the effects produced depending on the concentration of H2O2 to which these cells are exposed. A biphasic response was observed in osteoarthritic cell lines. A significant stimulation on hyaluronic acid (HA) synthesis in the presence of low concentrations of H2O2 (< 10 µmol/L) occurred, whereas an inhibitory effect of synthesis was noted at higher concentrations (> 10 µmol/L) [22]. During the course of OA, the synovial fluid undergoes degradation manifesting as a decrease in the amount and the average molecular weight of HA, which is correlated with pain and dysfunction[23, 24]. Notwithstanding the in vitro findings cannot be extrapolated to the human clinical situation, high molecular weight HA is an alternative treatment for knee OA that articular injection of HA acts to restore intraarticular lubrication, consequently improving joint biomechanics[25, 26]. Recent meta-analysis also suggests that HA injections are a safe and effective alternative in treating patients with symptomatic knee OA[25–29].
When present in excess, H2O2 is thought to be a key mediator in various processes that damage cells. Due to such reactive oxygen species (ROS) attacks, as growing evidence suggests the serious damage together with the alterations caused to physiological cell signalling lie in the core of the etiology and pathogenesis of several age-related diseases [30–32]. In OA, the augmented production of ROS and the significant reduction of antioxidant enzymes have lately been reported [33–37]. Accordingly, ROS may be an important contributor to the development of OA and a major risk factor in this degenerative disease besides aging[38].
In our PPI network constructed based on the DEGs, top 10 hub genes were identified: CYR61, PENK, GOLM1, DUSP1, ATF3, STC2, FOSB, PRSS23, TF, and TNC. Among them, CYR61, with the highest node degree as shown in Fig. 6 has gained our further focus. Cysteine-rich protein 61 (Cyr61/CCN1) is a product of an immediate early gene and a component of the extracellular matrix, playing a role in endothelial cell proliferation, differentiation, adhesion and migration[39, 40]. Despite this, previous studies discovered that during the progression of rheumatoid arthritis (RA), Cyr61 dramatically enhanced interleukin-17 (IL-17) expression in fibroblast-like synoviocytes and promoted the production of the pro-inflammatory cytokine IL-1β[41, 42]. Recently, Liu et al found that Cyr61 stimulated VEGF expression in osteoblasts and endothelial progenitor cells (EPCs)-primed angiogenesis, whereas Cyr61 knockdown inhibited angiogenesis in both in vitro and in vivo models. Furthermore, suppression of CYR61 ameliorated articular swelling and cartilage erosion in the ankle joint of murine collagen-induced arthritis (CIA) model[43]. Lin et al reported that Cyr61 produced from RA fibroblast-like synoviocytes (RA-FLS) initiated proliferation and migration of FLS, regulated IL-6 production in FLS, and induced differentiation of Th17 cell[44]. Another recent study suggested that Cyr61 promoted the production of IL-1β in FLS and induced osteoclastogenesis in RA[42]. In addition, FLS treated with Cyr61 increased IL-8 production and neutrophil influx in RA[45]. Together, these studies indicate that Cyr61 acts as a nodal effector molecule in the pathogenesis and progression of RA. As for its role in OA, Cyr61 has been reported to induce chondrogenesis and angiogenesis during embryogenesis of mice, and regulate chondrocyte maturation during cartilage development[46, 47]. But currently, the role of Cyr61 in OA synoviocytes has not been studied clearly yet. OA is generally considered as a form of chronic systemic low-grade inflammation, but in some cases the intensification of inflammatory lesions resembles changes seen in RA, which even hinders the differential diagnosis by means of imaging examinations[48–51]. This may have significant clinical implications. Therefore, the function of Cyr61 in OA might be somewhat alike but mostly different from that in RA, which needs further experimental study to confirm and analysis. The present study demonstrated that Cyr61 was differentially expressed in patients with OA and in normal group(as shown in Fig. 7), which might represent a starting point for subsequent researches into Cyr61 in OA synovium.
Compared with other studies of OA, the innovativeness of our present study is that GSEA was first used to conduct GO and KEGG analysis to explore the functions and pathways of DEGs. Moreover, the robust DEGs based on integrated bioinformatics analysis including GO and KEGG pathway enrichment, PPI network, and module analysis, may provide reliable molecular biomarkers from a novel perspectives to help us gain a better understanding of the disease pathogenesis. However, there existed several limitations. We need further study on large sample size to validate the results, and molecular experiments are required, especially for Cyr61.
In conclusion, DEGs associated with OA were screened through the GEO database, and integrated bioinformatics analysis was performed. As a result, a total of 300 DEGs and 10 hub genes were picked out, and we chose cellular response to H2O2 of synoviocytes and CYR61 to further discuss. Based on the effect of H2O2 a treatment option for knee OA patients is intra-articular injection of HA. But whether Cyr61 and its related pathways may be a candidate drug target for the treatment of OA, and its underlying mechanisms need further investigation.