Fibroblast growth factor 2: A new key player in osteoarthritis

Abstract

Vincent, Saklatvala, and colleagues have taken an interesting approach over the last 5 years by defining molecules that may be of importance in osteoarthritis (OA) through the investigation of proteins that are modulated following damage to articular cartilage. This approach has now paid off in a major way with the discovery that fibroblast growth factor 2 (FGF-2) is a key chondroprotective factor in an OA model, as reported by Chia et al (1) in this issue of Arthritis & Rheumatism. The authors made their original observations regarding FGF-2 through the investigation of ERK activation following explantation or explantation and subsequent cutting of cartilage (2). They showed that as rapidly as 30 minutes after damage, FGF-2 was liberated into the culture medium of cartilage, from which it could be purified. Later studies revealed that FGF-2 is found in a pericellular perlecan-bound pool in normal articular cartilage. Some FGF-2 was hypothesized to be liberated from this pool following mechanical loading, allowing it to stimulate ERK signaling as well as induction of certain matrix metalloproteinase (MMP) genes and tissue inhibitor of metalloproteinases 1 (TIMP-1) (3,4). Chia et al have now pursued these observations in an in vivo model of OA. Suppression of aggrecan degradation is regarded as a major goal in the prevention of cartilage destruction and loss of function in OA (for review, see ref. 5). Elegant studies in mouse models of arthritis have revealed that ADAMTS-5 is the key enzyme in aggrecan breakdown in this context (6,7). Moreover, biochemical analysis has revealed that human ADAMTS-5 has 1,000fold greater activity on aggrecan than does ADAMTS-4 (8). Studies of the regulation of aggrecanases have shown that the cytokine interleukin-1 (IL-1) induces ADAMTS-5 to a lesser degree than ADAMTS-4, but given the disparity in aggrecanase activity, this may not reflect a lesser importance for ADAMTS-5 (5). Although metalloproteinase expression levels in end-stage OA showed that ADAMTS-5 was among the proteinases that are repressed at this time point (9), recent genome-wide microarray studies of a surgically induced OA in the rat have revealed that Adamts5 is upregulated 4 weeks after surgery (10). These data suggest that ADAMTS-5 is likely to be important in early aggrecan cleavage during the development of OA, but may be repressed, perhaps through a feedback mechanism, during late stages of OA. Following their previous observations regarding FGF-2 up-regulation in damaged cartilage, Chia et al explored the effect of Fgf2 gene deletion on cartilage structure and on the development of OA-like characteristics in a surgically induced model of OA. They showed first that Fgf2 mice had increased development of OA with age as compared with wild-type (WT) mice, with marked effects at ages 6 and 9 months. At age 3 months, no differences were observed between genotypes in terms of cartilage breakdown or expression levels of several key genes, including aggrecan, type II collagen, and Adamts5. The authors did not examine the levels of these genes at ages 6 or 9 months, when one might expect to see an up-regulation of Adamts5 expression and function. Surgical destabilization of the knee was performed at age 10–12 weeks in both WT and Fgf2 mice, and the authors demonstrated that Fgf2 mice developed a more severe OA phenotype that was detectable 2, 4, and 8 weeks after surgery. Both cartilage thickness and proteoglycan content (estimated by Safranin O staining) were assessed. Of interest is the fact that Jelena Gavrilovic, PhD: University of East Anglia, Norwich, UK. Address correspondence and reprint requests to Jelena Gavrilovic, PhD, Cellular Protease Group, School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, UK. E-mail: J.Gavrilovic@uea.ac.uk. Submitted for publication March 6, 2009; accepted in revised form April 7, 2009. Arthritis & Rheumatism