Bone marrow stromal cells enhance prostate cancer cell invasion through type I collagen in an MMP-12 dependent manner.

At the cellular level, the process of bone metastasis involves many steps. Circulating cancer cells enter the marrow, proliferate, induce neovascularization, and ultimately expand into a clinically detectable, often symptomatic, metastatic deposit. Although the initial establishment and later expansion of the metastatic deposit in bone require tumor cells to possess invasive capability, the exact proteases responsible for this phenotype are not well known. The objective of our study was to take an unbiased approach to determine which proteases were expressed and functional during the initial interactions between prostate cancer cells and bone marrow stromal (BMS) cells. We found that the combination of human prostate cancer PC3 and BMS cells stimulates the invasive ability of cancer cells through type I collagen. The use of inhibitors for each of the major protease families indicated that 1 or more MMPs was/were responsible for the BMS-induced invasion. Gene profiling and semiquantitative RT-PCR analysis revealed an increased expression of several MMP genes because of PC3/BMS cell interaction. However, only MMP-12 showed an increase in protein expression. Downregulation of MMP-12 expression in PC3 cells by siRNA inhibited the enhanced invasion induced by PC3/BMS cell interaction. In vivo, MMP-12 was found to be primarily expressed by prostate cancer cells growing in bone. Our data suggest that BMS cells induce MMP-12 expression in prostate cancer cells, which results in invasive cells capable of degradation of type I collagen.

Host matrix metalloproteinase-9 contributes to tumor vascularization without affecting tumor growth in a model of prostate cancer bone metastasis.

Matrix metalloproteinases (MMPs) have been associated with initiation, progression and vascularization of a number of tumors. However, clinical trials using MMP inhibitors failed to meet expectations. Previously, we demonstrated the potential importance of MMP-9 activity in experimental prostate cancer bone tumor tissue. However, the particular roles of host- and tumor-derived MMP-9 remains to be defined. Herein, we examined the role of host MMP-9 in subcutaneous and intraosseous growth of the human androgen independent prostate cancer cell line PC3 in MMP-9 deficient mice. In the subcutaneous model, the tumor incidence in the control (RAG-1(ko/ko)) and experimental (RAG-1(ko/ko) /MMP-9(ko/ko)) group was 100%, with similar tumor growth kinetics and microvascular densities. In the intraosseous tumor model, the tumor incidence was higher in RAG-1(ko/ko) /MMP-9(ko/ko mice than in RAG-1(ko/ko) mice (67% and 39%, respectively), though no statistical differences were found. The intraosseous tumor areas were similar in both groups, and the number of tumor-associated osteoclasts did not differ significantly. However, the microvascular density of intraosseous tumors was higher in RAG-1(ko/ko) than in RAG-1(ko/ko)/MMP-9(ko/ko) mice, though no changes in tumor growth could be detected. In an in vitro assay, we found that bone marrow (BM) cells increased the invasiveness of PC3 cells, and that this enhancement was independent of MMP-9 expression by marrow cells. Our results with the RAG-1 model suggest that host-derived MMP-9 is neither necessary nor sufficient for subcutaneous or intraosseous PC3 tumor growth, osteoclastic response, or in vitro invasiveness of tumor cells.