Cancer cell metastasis to bone is a frequent occurrence with many common -human malignancies, including carcinomas of the prostate and breast Bone metastasis may be promoted by the phenotypic responses of cancer cells to factors which are present within the bone microenvironment This thesis describes in vivo and in vitro efforts to test the hypothesis that bone matrix proteins promote the growth and adhesiveness of metastatic cells. An animal model was developed to examine the growth properties of rat mammary cancer cells (Walker 256, W256) in metastatic target organs (Chapter 1). Rats were injected intramuscularly with W256 cells, which formed spontaneous metastases in the skeleton, liver, lungs, and kidneys within 7-10 days. Skeletal metastasis was associated with a progressive loss of trabecular bone. Within the skeleton, W256 cells which were immediately adjacent to trabecular bone had a 30% greater growth rate than did W256 cells located >50 μm from bone surfaces (p<0.05). These data established the first animal model of spontaneous bone metastasis, and provided preliminary evidence that bone matrix is a possible mitogen for metastatic cancer cells in vivo.

The mitogenic activity of bone matrix may be enhanced by osteoclastic bone resorption (1). To test this hypothesis, bone resorption was stimulated by injecting rats with Leydig tumor cells (Chapter 2). Leydig tumor burden was associated with decreased trabecular bone volume, fewer osteoblasts, and more osteoclasts. Leydig tumor burden was associated with a 56% increase in the growth rate of W256 cells in the skeleton (p<0.05), while W256 cell growth rate in the liver, lungs, and kidneys were not affected. The selective growth promotion of W256 cells in bone suggests the existence of a mitogen for W256 cells which is released from bone matrix during resorption.

Previous in vitro data suggest that this putative bone-derived mitogen may be transforming growth factor {$ (TGF-/3) (1), a factor which can stimulate oncogene expression. Media conditioned by resorbing fetal rat calvarial bones induced a rapid increase in W256 cell expression of c-myc oncogene mRNA and of nuclear c-myc protein (Chapter 3). Purified TGF-β also stimulated c-myc mRNA expression, suggesting that the mitogenic stimulation of W256 cells by bone-derived conditioned media may involve a signaling pathway which is also utilized during TGF-β-induced growth stimulation.

Bone resorption may be a sufficient but unnecessary growth stimulus for metastatic cells in vivo. To test this hypothesis, bone resorption in rats was inhibited in vivo with the bisphosphonate APD prior to their inoculation with W256 cells (Chapter 4). APD blocked the pathologic bone resorption normally associated with W256 tumor burden. However, the skeletal tumor burden in APD-treated rats was 2.6-fold greater than in untreated rats. Tumor burden in other organs was unaffected by APD. W2S6 cells in the skeletons of APD-treated rats had a 55% greater growth rate than did W256 cells in the skeletons of controls. W256 cells located adjacent to trabecular bone had greater growth rates than did W256 cells >50μm from bone, irrespective of APD treatment (p<0.05). These data suggest that bone matrix can exert a mitogenic influence on cancer cells without undergoing resorption.

Bone-derived factors are also capable of promoting adhesion (2), and the adhesion of metastatic cells to bone matrix could promote their skeletal localization. An in vitro model of bone matrix was developed to identify potential ligands which could support the adhesion of metastatic cells (Chapters 5 and 6). The extracellular matrix deposited by human U20S osteosarcoma cells supported the rapid adhesion of human PC-3 prostatic carcinoma cells. Antibodies directed against the integrin-lype collagen receptor α2β1 inhibited PC-3 cell adhesion to U20S matrix and to purified type I collagen, as did a collagen-derived peptide sequence (DGEA) which inhibits α2β1 function. TGF-β treatment caused a selective increase in α2β1 integrin expression, which was effected by increased rates of de novo synthesis of both α2 and β1 integrin subunits. The induction of α2β1 was associated with similar increases in the adhesion of PC-3 cells to U2OS matrix and to purified type I collagen. These data suggest that type I collagen may represent an important bone-derived adhesive ligand for metastatic cells. The abundant expression of type I collagen and of TGF-β in bone may promote bone metastasis by promoting the adhesion of metastatic cancer cells to bone matrix.