Deleterious interaction between BMP2 and TGF-beta signaling in articular chondrocytes: harmful role for BMP2 in OA?

Abstract

Purpose: TGF-beta signaling via Smad2/3 is crucial for cartilage maintenance. Smad2/3 maintains a healthy chondrocyte phenotype by inhibiting chondrocyte hypertrophy. We have previously shown that loss of Smad2/3 signaling is a hallmark of both ageing and OA. BMP2, a TGF beta family member, signals via Smad1/5/8 instead of Smad2/3. During OA we found that BMP2 is elevated near lesions. We have investigated whether BMP2, and as a consequence Smad1/5/8 signaling, affects Smad2/3 signaling and what are the biological consequences of this. Methods: Primary bovine chondrocytes were exposed to BMP2 followed by Western blotting for Smad2/3P after 60, 180 and 420 minutes. To study whether the observed effects were reproducible in vivo we injected Ad-BMP2 i.a. into murine knee joints; Ad-luc was used as a control. We isolated knee joints after 3 days for immunohistochemistry on Smad2/3P. To address biological relevance of decreased Smad2/3P, human OA cartilage explants obtained from patients undergoing joint replacement surgery after informed consent were exposed to the Smad2/3P inhibitor SB-505124. After 48 hours RNA was isolated to investigate MMP13 levels as a marker for hypertrophy and OA marker. In addition, we exposed primary bovine chondrocytes to BMP2 and evaluated whether that had an effect on MMP13 levels as well.As we observed that BMP2 reduced Smad2/3P signaling and stimulated MMP13 expression we wanted to investigate whether this effect could be reversed by extra stimulation of the Smad2/3 route. Therefore chondrocytes were exposed to a combination of BMP2 and TGF-beta. Moreover, human OA cartilage explants were exposed to TGF-beta to investigate whether this reduced MMP13 expression. Results: Exposing primary bovine chondrocytes to 15 or 50 ng BMP2 led to a decrease in Smad2/3 phosphorylation on Western Blot already after 60 minutes and that was sustained until at least 420 minutes compared to non-stimulated controls. We confirmed this effect in vivo as over expression of BMP2 in murine knee joints led to decreased Smad2/3P positive cells in both patella and tibia by 55% and 43% respectively. To investigate the functional consequence of decreased Smad2/3 signaling we cultured human OA cartilage explants in the presence of SB-505124. This led to a significant increase in MMP13 expression of 13,6 fold (p<0.02). A similar increase in MMP13 expression was also observed in cultures of primary bovine chondrocytes 48 hours after stimulation with BMP2. Interestingly, the effect of either 15 or 50 ng BMP2 could be prevented by addition of TGF-beta, which prevented the decrease in Smad2/3P levels, but only when a dose of 10 ng/ml TGF-beta was used. A lower dose of 0.5 ng/ml TGF-beta was insufficient to overcome the BMP2-induced Smad2/3P decrease. This indicates that TGF-beta can overcome the BMP2 effect in a dose-dependent manner. Even in human OA cartilage that had spontaneously reduced Smad2/3P levels addition of 10ng/ml TGF-beta could still decrease MMP13 expression by 6.6 fold (p<0.005). Conclusions: Our data confirm that BMP2 down regulates Smad2/3 signaling in articular chondrocytes both in vitro and in vivo. A decreased Smad2/3 signaling in turn results in up regulated levels of MMP13 expression suggesting that BMP2 up regulation, as found adjacent to OA lesions, is deleterious for articular cartilage. Moreover, we show that this negative effect of BMP2 can be overcome by addition of TGF-beta, thereby raising levels of Smad2/3P. Even in human OA cartilage that has naturally high levels of MMP13 and low levels of Smad2/3P, TGF-beta addition can still reduce MMP13 expression levels. Overall our data suggest that BMP2 induces OA-like effects in articular cartilage, down regulating Smad2/3P and up regulating MMP13, which can be overcome by enhanced TGF-beta signaling.