Is cancer stem cell a cell, or a multicellular unit capable of inducing angiogenesis?

Cancer stem cells are presently viewed as carriers of the growth initiating potential, repopulation capability and drug resistance in tumors. However, many of these fundamental properties of cancer are host-related, modified by cell-cell interactions and/or dependent on angiogenesis. Indeed, it is well established that co-injection of cancer cells with their irradiated (mitotically dead) counterparts, or with Matrigel can significantly increase their tumor forming capacity (i.e. the attribute presently associated with cancer cell 'stemness'). Similarly transfection of angiogenic factors (e.g., VEGF/VPF) can promote such capacity in certain cell lines. Moreover, injection site (e.g., orthotopic vs. ectopic) may significantly modulate tumor take in experimental settings. These observations cannot be reconciled with the paradigm that tumor initiation potential is a fixed, constitutive and cell autonomous feature of a subset of cancer cells expressing stem cell markers (e.g., CD133, Sca1 and other). Instead, it is proposed here that 'stemness' in cancer (perhaps unlike in normal self renewing tissues), rather then being assigned to a particular readily identifiable cell subset, could be a property of interactive clusters of cancer cells (perhaps including, but not limited to cells with stem cell markers). Such 'multicellular units' would become equipped with properties experimentally perceived as 'stemness', i.e. the capacity to initiate tumor growth, when they express the capacity to induce angiogenesis. It is also postulated here that, while the pursuit of subsets of cancer cells harbouring stem cell markers has been fascinating and revealing, due to aforementioned limitations of the present stem cell concept and presumed intractability (e.g., mutability) of such cells, further therapeutic promise may reside in a better definition of 'multicellular angiogenic cancer stem units'.