Persistence of human vascular endothelium in experimental human prostate cancer bone tumors.

Using the SCID-human model, we recently found that human circulating prostate cancer cells formed tumors in human bone but not mouse bone (Nemeth et al. Cancer Res 1999; 59: 1987-93). It is possible that this tissue preference was mediated by interaction between human tumor cells and human endothelial cells within the implanted bone tissue. We sought to determine the relative amounts of human and mouse vasculature within human bone implants and resulting prostate cancer bone tumors in the SCID-human model. Paraffin sections of plain bone implants or PC3 or LNCaP human bone tumors were double stained for factor VIII (all vessels) and human CD31 (human vessels) followed by fluorescent secondary reagents. At 4 weeks post implantation (when cancer cells are typically introduced), the vasculature within human bone fragments remained primarily human (84.5%), and this pattern persisted to at least 10 weeks (91.6% human). Injection of PC3 cells into the bone resulted in an increase in mouse-derived vessels, however the majority (58%) of the vessels remained human even after the formation of large bone tumors. LNCaP bone tumors were highly angiogenic, and there was a sharp decline in the proportion of vessels which were antigenically human (36.8%), suggesting recruitment of mouse endothelial cells during the angiogenic process. Nonetheless, the persistence of human vasculature suggests the SCID-human model can be used to study the interaction between bone-seeking tumor cells, such as prostate cancer, and human bone endothelium in vivo, and to test potential therapeutic strategies which may depend on the presence of human vessels.