SKELETAL ALKALINE-PHOSPHATASE ACTIVITY IS PRIMARILY RELEASED FROM HUMAN OSTEOBLASTS IN AN INSOLUBLE FORM, AND THE NET RELEASE IS INHIBITED BY CALCIUM AND SKELETAL GROWTH-FACTORS

Skeletal alkaline phosphatase (ALP) is anchored to membrane inositol-p hosphate on the outer surface of osteoblasts. Although skeletal ALP ac tivity in serum is, essentially, all in an anchorless (soluble) form, in vitro studies indicate that ALP can be released in either an anchor less, soluble form (e.g., by a phospholipase) or an anchor-intact, ins oluble form (e.g., by vesicle exocytosis). The current studies were in tended to define the contributions of each of these putative processes of ALP release and to assess the significance of regulation by calciu m (Ca) and skeletal effecters. ALP activity was measured in serum-free medium from replicate cultures of human osteosarcoma (SaOS-2) cells a nd normal human bone cells. Temperature-sensitive phase distribution ( in Triton X-114) allowed separation of soluble from insoluble ALP acti vity. Our studies revealed that most of the ALP activity released from SaOS-2 cells was in an insoluble form (78% +/- 8%), a percentage that was constant between 2 and 96 hours. A similar result was seen for no rmal human bone cells. Calcium had a negative, biphasic dose-dependent effect on net release of ALP activity: r = -0.85, P < 0.001 at 24 hou rs, with K-Iapparent values for biphasic inhibition of 20 and 300 mu m ol/l Ca. Of the skeletal effecters tested, insulin-like growth factor- II (IGF-II had the greatest effect, decreasing the net release of ALP activity in a dose-dependent manner (r = -0.82, P < 0.005). Neither Ca nor IGF-II affected the distribution of soluble/insoluble ALP activit y by more than 9%. IGF-II had no effect on extracellular ALP stability , but the addition of Ca to Ca-free cultures resulted in parallel loss es of extracellular ALP activity and ALP immunoreactive protein (P < 0 .001 for each). A similar effect was seen when Ca was added to Ca-free , cell-free, conditioned medium, but not when Ca was added to purified ALP, which is consistent with the general hypothesis that a Ca-depend ent protease might be present in the cell-conditioned medium. Together , these data suggest that most of the ALP activity released from osteo blasts is insoluble (and, presumably, anchorless), net release of ALP activity is negatively regulated by Ca and skeletal growth factors, th e effect of Ca may reflect Ca-dependent protease activity, and an exog enous (e.g., serum) phospholipase may be responsible for releasing ALP from its insoluble anchor.