We have demonstrated that canonical and non-canonical Wnt signaling occurs in myeloma cells (Qiang et al., 2005) and overexpression of Wnt3a in myeloma cells inhibits the osteolytic phenotype and also tumor growth in vivo (Qiang et al Blood, Abstract #3420, 2006). To further investigate the mechanisms that contribute to this process we have expanded our in vivo data by showing that while H929 cells stably expressing Wnt3a (H929/W3a) leads to reduced tumor growth in the in-vivo SCID-hu bone graft model compared with H929 vector alone transfected control cells (H929/EV), there was no significant difference in the subcutaneous growth of the two cell lines in SCID mice. Taken together these data suggests that alteration of the human bone marrow microenvironment is central to Wnt-mediated reduction in tumor growth in bone. We next employed an in-vitro co-culture model in which the mouse osteoprogenitor cell line, C2C12, and human osteoblast cell line, Saos-2 were co-cultured with either H929/Wnt3 or H929/EV cells. QPCR analysis demonstrated that osteoprotegerin (OPG) mRNA expression (relative OPG mRNA to GAPDH) in C2C12 cells co-cultured with H929/W3a was significantly elevated compared with H929/EV (mean±SD: 14.34±0.97 vs 8.43±0.16; P<0.001). ELISA analysis showed that OPG protein levels in the cell culture supernatant were also significantly higher (71.02 ± 6.178 vs 0 pg/ml; P<0.001). Similar results in OPG mRNA and protein levels were observed in Saos-2 cells co-cultured with H929/W3a relative to H929/EV. Furthermore, treatment of C2C12 cells with recombinant Wnt3a protein induced both OPG mRNA (48.1 ±1.2 vs 1.0±0.5; P<0.001) and protein levels (1767.03 ± 44.8 vs 1.11 ± 0.03 p< 0.0001) compared with vehicle alone. These results suggest that forced expression of a canonical Wnt ligand by MM cells might promote OPG transcription in osteoblast progenitors in-vivo. To further confirm the role of Wnt signaling in regulation of OPG and RANKL transcription, we produced C2C12 cells that stably express Dkk1. These clones showed a significant inhibition of Wnt3a induced OPG mRNA (22.2± 2.3 vs 1.7±0.35; p<0.001) and protein (73.3 ± 18.0 vs. 0 pg/ml; p<0.01) compared with vector control. In contrast, RANKL mRNA (5.1±0.9 vs 1.0± 0.5, p<0.01) and protein (9.3±3.8 vs. 0 pg/ml; p<0.01) were increased in Dkk1 expressing clones compared with control. Moreover, supernatant from C2C12 clones stably expressing a DN-beta-catenin (DNBC/C2C12) contained a significantly higher level of RANKL (17.3± 3.5 pg/ml vs. 0±0; P<0.001) and a dramatically lower level of OPG protein (0±0 vs. 431.186 pg/ml; P<0.001) compared with control. Finally, the numbers of multinuclear TRAP-positive osteoclasts were significantly more abundant in culture containing supernatant from DNBC/C2C12 than that from vector control, while Wnt3a exposure had no effect on osteoclast formation in-vitro. Taken together, these data suggest that Wnt ligand-mediated inhibition of myeloma cell growth, and inhibition of osteolytic lesions, in-vivo may result from upregulation of OPG and loss of RANKL in osteoblast progenitors, which subsequently diminishes osteoclast formation. Results of these studies provide new insights into mechanism by which Wnts may serve as an important indirect regulator of myeloma growth and osteoclast formation, and as such, targeting Wnt signaling may be an new therapeutic strategy for controlling myeloma growth and associated bone disease.

Author notes

Disclosure: Consultancy: Millennium, Novartis, Cephalon. Research Funding: Millennium, Novartis, Zymogenetics. Honoraria Information: I3CME. Financial Information: Novartis.