Search

Bone metastases are often associated with osteolysis and subsequent pathological fractures. To determine if metastatic human cancer cells can directly degrade non-mineralized and mineralized bone, we used prostate PC3 adenocarcinoma cell lines, which were originally established from skeletal metastases. We show that PC3 cells and their conditioned medium degraded non-mineralized, osteoid-like radiolabelled extracellular matrices from human Saos2 and U2OS osteoblast-like cells. These cells also directly degraded mineralized bone by inducing (45)Ca release from rat fetal calvariae and forming resorption pits on bone slices, an effect increased by transforming growth factor-beta(1). A role for matrix metalloproteinases in degradation was shown by: (1) stimulation by the phorbol ester TPA of PC3-induced matrix degradation and release of matrix metalloproteinase activity; (2) abrogation of matrix degradation by 1,10-phenanthroline, a metalloproteinase inhibitor, and (3) degradation of purified type I collagen by PC3 cells and their conditioned medium. We demonstrate that human prostate cancer cells can directly degrade bone-related matrices and that matrix metalloproteinases have a role in this process., (Copyright 2000 S. Karger AG, Basel.)

Experiments in vivo have demonstrated that endothelial cell injury promotes the local arrest of circulating, intravascular cancer cells and the subsequent formation of metastatic tumors. The experiments described here were performed to test the hypothesis that injury of the endothelium also causes damage to the adjacent vascular basement membrane, which in turn facilitates the passage of cancer cells across the vessel wall. Confluent monolayers of bovine pulmonary artery endothelial cells were incubated with 3H-2-deoxyglucose or 3H-proline to label the endothelial cells or the basement membrane, respectively. After adding H2O2 to these cultures, damage of the endothelium and basement membrane was detected by release of the isotopes into the culture medium. The kinetics and magnitude of basement membrane degradation correlated with the damage to the endothelial cells. Evidence for involvement of endothelial proteases in basement membrane injury included identification of a 63-kD gelatinase in the culture medium, inhibition of injury by protease inhibitors and the inability of H2O2 to cause 3H-proline release when applied directly to basement membranes. Scanning electron microscopy demonstrated that a greater number of A549 lung adenocarcinoma cells attached to the basement membrane and endothelium at points of endothelial retraction. However, this was not due to an increase in the adhesive properties of the basement membrane. The media from injured endothelial cultures stimulated the motility of A549 cells in a Boyden chamber assay. Furthermore, in a 24-hour invasion assay, a greater number of A549 cells migrated through injured basement membranes than through control membranes. We conclude that endothelial cell injury can cause enzymatic damage to the underlying basement membrane and postulate that this can facilitate the transvascular passage of cancer cells in vivo.

Adhesion of tumor cells to vascular endothelial surfaces is one of the key steps in metastatic dissemination. Several factors are believed to be implicated in the regulation of the adhesive properties of tumor cells. We show that the adhesion of five different tumor cell lines, all of them of human origin, to human umbilical vascular endothelial cells (ECs) significantly increases following pretreatment of ECs with the cytokines interleukin-1 and tumor necrosis factor, whereas tumor cell/EC interactions remained unchanged after incubation with interferon-gamma. Significant augmentation in tumor cell adhesion was also observed when ECs were treated with the lipoxygenase inhibitors salicylate and the compound BW755C. In all cases, increased tumor cell adhesion was concomitant with significant decreases in the EC levels of linoleic acid, lipoxygenase-derived metabolite 13-hydroxy-octadecadienoic acid (13-HODE). On the contrary, pretreatment of the EC monolayers with aspirin did not result in any changes towards tumor cell adhesion. These results suggest that tumor cell/EC interaction is modulated, at least in part, by intracellular levels of 13-HODE and is independent of prostacyclin (PGI2) production by the ECs.

Bone is a major source of transforming growth factor-beta (TGF-beta), and a preferred target organ for metastasis of the rat Walker carcinosarcoma 256 (W256). Since chemotactic mechanisms may contribute to metastatic site specificity, we have tested the hypothesis that TGF-beta is a chemoattractant for these cancer cells. Purified platelet-derived TGF-beta elicited dose-dependent migration of W256 cells in the Boyden chamber assay with half-maximal responses (ED50) elicited by 0.12 +/- 0.01 ng TGF-beta/ml. Checkerboard analysis confirmed dependence of the response upon a concentration gradient. Conditioned media from organ cultures of bone contained TGF-beta and chemotactic activity in proportion to the extent of bone resorption. The chemotactic activity in conditioned bone culture medium and that of the purified platelet-derived TGF-beta were both inhibited after incubation with anti-TGF-beta 1. We conclude that TGF-beta, released from resorbing bone, can influence the migratory behavior of the osteotropic W256 cell line.

Recent reports indicate that a wide range of mammalian cells exhibit chemotactic behavior. Unlike the neutrophil, which has been most extensively studied, many of these cells are capable of cell division and passage in vivo and in tissue culture. In this paper we describe studies on the Walker carcinosarcoma 256, a rat mammary tumor which exhibits chemotactic responses to factors in the media of cultured resorbing bone and to a factor generated by proteolysis of the fifth component of serum complement (C5). We demonstrate here that long-term passage in vitro, passage in vivo, the phase of growth, and the heterogeneous nature of this tumor can have a significant effect on the chemotactic behavior of Walker carcinosarcoma cells.

The formation of metastases can be induced at sites injected with chemotactic factors in animals with circulating, chemotactically responsive tumor cells. This study was done to examine the possibilities that this phenomenon is due to enhanced tumor localization or stimulation of tumor growth. Walker carcinosarcoma cells labelled with 125I-iododeoxyuridine were injected into the tail veins of rats subsequently given intraperitoneal injections of chemotactic factor. 24 h later there was a 2.7- to 3.5-fold increase in the number of labelled cells in the mesenteries of animals given chemotactic factors compared to controls (p less than 0.001). The distribution of cells in other viscera was not affected by the injection. No effects were observed on the kinetics of tumor growth in vitro when cultures were supplemented with these chemotactic factors or with peritoneal lavage fluids from animals previously given intraperitoneal injections of chemotactic factor. Thus, in this model, the formation of metastases can be related to the arrest of circulating tumor cells in response to local chemotactic stimuli but the promotion of tumor growth has not been demonstrated.

The lung, a frequent site of cancer metastases, is also a susceptible target in several models of endothelial injury. In previous studies we have demonstrated that such injury, induced by bleomycin or by exposure to high concentrations of atmospheric oxygen, can facilitate the localization and metastasis of circulating tumor cells. Here we have tested the hypothesis that neutrophil-mediated injury to pulmonary endothelium has a similar effect. In Sprague-Dawley rats, intravenous injection of cobra venom factor resulted in complement activation, rapid sequestration of neutrophils in the lung, and endothelial damage, demonstrated by morphology, and by increased protein content and leakage of intravenous 125I-albumin into bronchoalveolar lavage fluids. When 125I-iododeoxyuridine-labeled Walker carcinosarcoma cells were injected intravenously during the period of endothelial injury, the pulmonary capillaries contained aggregates of neutrophils and tumor cells 2 h later, and there was a 3-fold increase in pulmonary tumor cell localization after 24 h in treated animals, compared to controls. Enhancement of tumor cell localization was prevented by pretreatment of the rats with catalase or by antineutrophil antisera. When animals were examined 2 weeks after cell injection, treatment groups had significantly more metastatic tumor nodules and a greater area of lung tissue involved by metastatic tumors. We conclude that neutrophil-mediated damage to the pulmonary endothelium can significantly increase the trapping of circulating tumor cells, and is likely to be clinically important since the large vascular bed of the lung is susceptible to host-mediated injury.