Objective: Subchondral bone sclerosis is a common feature of osteoarthritis (OA), but the mechanisms responsible for this condition remain unresolved. We investigated the role of insulin-like growth factor 1 (IGF-1) and urokinase plasminogen activator (uPA) in human osteoblasts from subchondral bone obtained from the tibial plateaus of OA patients and normal individuals.
Methods: Primary in vitro osteoblasts were prepared from subchondral bone specimens obtained from OA patients at surgery and from normal individuals at autopsy. Levels of uPA and PA inhibitor 1 (PAI-1) levels were determined under basal conditions and after IGF-1 stimulation in conditioned media from osteoblasts by enzyme-linked immunosorbent assay. The activity of uPA was evaluated by specific substrate hydrolysis and zymography under basal conditions and after plasminogen stimulation, in the presence and absence of added IGF-1. Plasmin activity was also evaluated by specific substrate hydrolysis.
Results: Levels of uPA released by OA osteoblasts were significantly higher than normal. Addition of IGF-1 to osteoblasts significantly reduced uPA protein levels only in OA patients (P < 0.05). In contrast, the addition of uPA to osteoblasts did not modify IGF-1 levels in either normal or OA osteoblasts. Basal uPA activity was higher in OA than in normal osteoblasts. Interestingly, IGF-1 enhanced basal uPA activity in OA specimens in a dose-dependent manner. Addition of plasminogen promoted uPA activity in both normal and OA osteoblasts via a positive feedback loop due to plasmin generation, since this activity was inhibited by both PAI-1 and alpha2-antiplasmin. Unexpectedly, incubation with IGF-1 inhibited this positive feedback of plasminogen-dependent uPA activity in OA osteoblasts, but not in normal osteoblasts, in a dose-dependent manner. Hence, normal osteoblasts were relatively insensitive to IGF-1, whereas the same treatment reduced both uPA levels and plasminogen-dependent uPA activity in OA osteoblasts while it increased basal uPA activity in OA osteoblasts. This could not be explained by PAI-1 protein levels, which were similar in normal and OA osteoblasts in the presence and absence of IGF-1. IGF-1 also reduced plasmin activity in OA osteoblasts while it did not modify this activity in normal osteoblasts.
Conclusion: These results suggest that in OA osteoblasts, the uPA/plasmin system functions normally, yet IGF-1 inhibits the positive feedback of plasmin on uPA activity. This inhibition may contribute to abnormal IGF-1- and uPA-dependent bone remodeling, ultimately leading to abnormal bone sclerosis in OA.