Interactions between insulin-like growth factor-I (IGF-I) and the system of plasminogen activators and their inhibitors in the control of IGF-binding protein-3 production and proteolysis in human osteosarcoma cells.

Limited proteolysis in vivo of insulin-like growth factor-binding protein-3 (IGFBP-3) by as yet unidentified serine proteases plays a key role in controlling the bioavailability of IGFBP-3-associated insulin-like growth factors (IGFs). Both the IGF system and the system of plasminogen activators (PAs) and their inhibitors (PAIs) are involved in bone remodeling, and plasmin has been shown to provoke dissociation of IGFBP-IGF complexes in cultured MG-63 human osteoblasts. The aim of this work was to investigate interactions between IGF-I and the PA/PAI system and their influence on IGFBP-3 production and proteolysis in this cell model. At confluency, MG-63 cells maintained for 3 days in serum-free medium constitutively secreted IGFBP-2 and small amounts of IGFBP-3 and IGFBP-4. As shown by Western ligand and immunoblot analyses of the culture medium and Northern blot analysis of IGFBP-3 messenger RNA, production of these IGFBPs, and of IGFBP-3 in particular, was dose dependently stimulated by the addition of 12.5-100 ng/ml recombinant human (rh) IGF-I. Increasing concentrations of plasminogen (0.05-3.5 micrograms/ml) added during the final 12 h of culture reduced the amounts of IGFBP detectable by Western ligand blotting, especially IGFBP-3. This reduction reflected proteolysis, as shown by immunoblotting, which revealed 30-, 20-, and 16-kilodalton fragments of IGFBP-3. In the presence of 25 ng/ml IGF-I, which stimulated IGFBP-3 production, proteolysis was reduced by approximately half. Incubation of glycosylated [125I]rh-IGFBP-3 as substrate in cell-free conditioned medium gave the same results. Addition of 50 ng/ml rhIGF-I to conditioned medium (to promote IGFBP-3-rhIGF-I complex formation) failed to diminish plasmin-induced proteolysis of IGFBP-3. Urokinase PA activity in the conditioned medium decreased significantly when the cells were cultured with rhIGF-I, indicating a direct action of IGF-I on urokinase PA and/or PAI production. Our results support the notion of a regulation loop whereby IGF-I controls its own bioavailability via its action on both IGFBP-3 production and the PA/PAI system, which regulates IGFBP-3 proteolysis. The proteolytic cleavages of IGFBP-3 caused by plasmin were the same as those caused in vivo by serine protease acting on this IGFBP.