Metastasis at the bone site, frequently found in many breast cancers, is a complex process where tissue destruction induced by the tumor cells is an integral part of tumor development. Since host cells and host factors actually cause the tissue destruction and bone loss, the interaction of host and tumor cells in the development of tumors is underscored in this tissue. We have developed a transformed mouse mammary epithelial cell line (r3T) that forms bone metastases in syngeneic mice: this model allows us to use genetically modified host mice to explore the role of various host proteins in the process of metastatic tumor growth in the bone. The secreted phosphoprotein osteopontin (OPN) enhances tumor growth in mice and is strongly associated with tumor progression in human cancer. OPN is also a major non-collagenous protein of bone where it is required for maximal bone resorption. OPN-deficient mice are resistant to several types of stimulated bone resorption, for example, the bone loss that occurs in response to ovariectomy. Here, we have utilized our new model of bone metastasis to evaluate the effect of host osteopontin on metastatic tumor growth and the resulting bone loss. We hypothesized that the OPN deficient mice would be resistant to the bone loss induced by tumor cell growth at the bone site, and would likely show reduced tumor burden as well. Like the MDA-MB-231 cells, r3T cells do not express RANKL, but do express PTHrP. When r3T cells were injected via the intracardiac route into both WT and OPN −/− mice the mice developed bone metastases and cachexia. Mice were sacrificed when moribund, and tumor burden and bone loss evaluated: the OPN −/− mice were sacrificed slightly earlier than their WT counterparts, indicating that host OPN has a suppressive effect on tumor development in the bone microenvironment. We developed and validated a real-time PCR assay for the presence of tumor cells in the bone, and used this assay to demonstrate that the size of the tumors in the bones of mice of the two genotypes did not differ significantly. Unexpectedly, analysis of bone loss by x-ray and micro CT indicated that the extent of bone loss in the proximal tibia was similar in both WT and OPN −/− mice bearing similar-sized tumors. In the tumor-bearing bones, PTHrP and RANKL expression were both elevated, and in bones with large tumors, the expression of these two genes was similar and independent of genotype. Taken together, these results indicate that stimulation of bone resorption by tumor cells is independent of osteopontin. OPN-deficient osteoclasts show reduced motility and impaired bone resorption in vitro, and our results suggest that tumor cells can override these deficiencies. The mechanism by which this happens is a subject of intense interest. Supported by NIH grant # DK67685.

[Proc Amer Assoc Cancer Res, Volume 46, 2005]