Prostate cancer metastasis is a unique disease. The propensity of prostate cancer to metastasize to bone and the prognostic significance of bone metastasis suggest that effective treatment of bone metastasis may provide clinical benefits. Both osteoblasts and osteoclasts have been shown to play a central role in the interactions between the metastatic prostate cancer cells and bone. Although most prostate cancer bone metastasis is osteoblastic, it remains unclear which cell type is initially involved in the process. Other components in the bone, such as the endothelium and stroma, may also play important roles in this process. The osteoblastic feature of prostate cancer bone metastasis has led to therapies focused on targeting osteoblast activity. Clinical trials targeting osteoblasts use radiopharmaceuticals (strontium-89 and samarium-153), the endothelin A receptor inhibitor atrasentan, or the vitamin D analog calcitriol. Agents that target osteoclasts include bisphosphonates; those that target endothelial cells include thalidomide and bevacizumab. Although these clinical trials for bone metastasis may provide effective treatments, novel concepts of how prostate cancer cells selectively metastasize to bone may advance our understanding and provide improved treatments for this difficult clinical problem. We propose a refined "seed" and "soil" view of metastasis. We speculate that in prostate cancer bone metastasis, the seed may comprise the so-called cancer stem cells, whereas the soil may comprise an onco-niche, ie, a unique microenvironment, that facilitate the growth and survival of cancer stem cells. If so, targeting cancer stem cells or the onco-niche may offer another way to improve treatment of prostate cancer bone metastasis.