We have read with interest the Review by Begg, Stewart and Vens (Strategies to improve radiotherapy with targeted drugs. Nature Rev. Cancer 11, 239–253 (2011))1, which included discussion on the potential benefits of modulating the tumour microenvironment to improve radiotherapy. In their Review, the authors state that, “vasculogenesis (vascular formation from circulating BMDCs (bone marrow-derived cells)) has been shown to be crucial for the growth of tumours that recur after radiotherapy”. We agree that new vessel formation is an important determinant of tumour regrowth after irradiation. However, we question the authors' conclusion that vasculogenesis is crucial in this context.

To our knowledge, there are no published reports showing greater recruitment and incorporation of non-irradiated endothelial progenitor (or precursor) cells (EPCs) — or any other BMDCs — into the vessels of tumours that are regrowing after local irradiation compared with their ionizing radiation-naive counterparts. In fact, three animal studies have directly tested this hypothesis either in irradiated tumours or using tumours implanted in pre-irradiated beds — situations in which the local endothelial cells that are capable of neovascularization have been substantially damaged by ionizing radiation. None of these studies found evidence for increased BMDC incorporation into the vessels of such tumours2,3,4. Moreover, one of these papers — overlooked by Begg and colleagues in their Review — emphasized that the recruitment of EPCs does not facilitate tumour regrowth after local irradiation4.

The discrepancy seems to be due to the initial, imprecise use of the term 'vasculogenesis' elsewhere2,3. Although studies have shown that BMDCs increasingly infiltrate into tumours after irradiation and facilitate relapse3,4,5, they do not incorporate into the vasculature. Thus, these BMDCs do not participate in vasculogenesis as defined by Begg and colleagues and in the literature in general1,6. Indeed, the BMDCs that are recruited by irradiated tumours are generally of myeloid lineage — including 'alternatively' activated macrophages and tyrosine protein kinase receptor (TIE2)-expressing monocytes (TEMs) — that facilitate post-irradiation tumour revascularization through paracrine mechanisms2,3,4,5. Distorting the definition of vasculogenesis not only confuses the literature but also misrepresents the biology — in this case by implying that the cells are doing something that they are not.

Interestingly, several recent discoveries might add new twists to the issue. It has been shown that a fraction of endothelial cells in growing tumours may be derived, through trans-differentiation, from cancer stem-like cells7,8,9, which may be more radioresistant than other neoplastic cells10,11. Alternatively, EPCs that are not of bone marrow origin might potentially be recruited to tumours from non-irradiated adjacent tissues and/or the circulation12,13. It is conceivable that all of these cells may be participating in vasculogenesis in tumours that recur after irradiation. But at this stage, this remains a hypothesis. Only when these or other proposed mechanisms of EPC differentiation and incorporation are proven in irradiated tumours should we conclude that vasculogenesis is crucial in relapsing tumours.