Metastasis-associated differences in gene expression in a murine model of osteosarcoma

Despite advances in the management of osteosarcoma (OSA) and other solid tu mors, the development of metastasis continues to be the most significant pr oblem and cause of death for cancer patients. To define genetic determinant s of pulmonary metastasis, we have applied cDNA microarrays to a recently d escribed murine model of OSA that is characterized by orthotopic tumor grow th, a period of minimal residual disease, spontaneous pulmonary metastasis, and cell line variants that differ in metastatic potential. Microarray ana lysis defined 53 genes (of 3166 unique cDNAs) that were differentially expr essed between the primary tumors of the more aggressive (K7M2) and less agg ressive (K12) OSA models. By review of the literature, these differentially expressed genes a ere assigned to six nonmutually exclusive metastasis-ass ociated categories (proliferation and apoptosis, motility and cytoskeleton, invasion, immune surveillance, adherence, and angiogenesis). Functional st udies to evaluate K7M2 and K12 for differences in each of these metastasis- associated processes revealed enhanced motility, adherence, and angiogenesi s in the more aggressive K7M2 model. For this reason, 10 of the 53 differen tially expressed genes that were assigned to the motility and cytoskeleton, adherence, and angiogenesis categories mere considered as most likely to d efine differences in the metastatic behavior of the two models. Ezrin, a ge ne not described previously in OSA, with functions in motility, invasion, a nd adherence, was 3-fold overexpressed in K7M2 compared with K12 by microar ray. Differential expression for RNA was confirmed by Northern analysis and for protein by immunostaining. Alterations in ezrin protein levels and con comitant cytoskeletal changes in our model confirmed predictions from the a rrays. The potential relevance of ezrin in OSA was suggested by its express ion in five of five human OSA cell lines. This work represents a rationale approach to the evaluation of microarray data and will be useful to identif y genes that may be causally associated with metastasis.