Figure 7

Hypothetical scheme of the mechanism by which Dkk1 promotes osteolytic lesions in vivo. A, The cellular differentiation of osteoprogenitors, which can be monitored by Osterix and ALP expression, is governed in part by Wnt3a and BMP2 signaling. BMP2, which is inhibited by Noggin, is a strong inducer of Osterix and ALP, relying on an indirect mechanism that involves p53 and Cbfa1 (not included). Wnt3a/β-catenin signaling, which is antagonized by Dkk1 (and Dkk2), can have opposing effects on osteoblast development depending on BMP2 signaling. In absence of BMP2, Wnt3a weakly promotes osteoblast differentiation through the β-catenin pathway. In presence of BMP2, Wnt3a inhibits osteoblast differentiation by down-regulating the BMP2-p53-Cbaf1-Osterix-ALP axis (thick line). The molecular mechanism of the crosstalk has not yet been elucidated. Our scheme postulates that the Dkk1-dependent inhibition of Wnt3a's positive effect on osteoprogenitors in vitro may be outweighed by the abrogation of Wnt3a's negative effect on BMP2 signaling, such that, on balance, Dkk1 enhances features of early osteoblastogenesis. B, The differentiation of osteoclasts, which is strictly dependent upon interaction of osteoclast precursors with osteoblasts, is mediated in part by signaling pathways that emanate from the binding of M-CSF and RANKL to c-Fms and RANK, respectively. Osteoblasts stimulated by BMP2 express high levels of M-CSF and RANKL (positive regulators of osteoclastogenesis) but low levels of OPG (negative regulator of osteoclastogenesis), resulting in a positive signal for osteoclast formation. Osteoblasts exposed to Wnt3a exhibit the opposite phenotype with regard to M-CSF, RANKL and OPG expression, resulting in a negative signal for osteoclast formation. Osteoclastogenesis is further dampened by Wnt3a's ability to inhibit BMP2 signaling, as indicated in panel A (thick line). Constitutive expression of Dkk1 (and Dkk2) by bone marrow-homing tumor cells (e.g., multiple myeloma) may enhance osteoclastogenesis in vivo by inhibiting Wnt3a, thus shifting the balance of Wnt3a/BMP2 co-signaling in favor of BMP2 signaling. The consequent increase in osteoclastogenesis promotes lytic bone lesions as long as the paracrine supply of Dkk1 by tumor (myeloma) cells is sustained.