Are tumours angiogenesis-dependent?

The final proof of principle that cancer patients can be effectively treated with angiogenesis inhibitors is eagerly awaited. Various preclinical in vivo experiments have proven that most tumours need new vessel formation in order to grow and to form metastases. First of all, tumours do not grow in avascular corneas until new blood vessels reach the implant. Secondly, the introduction of only one angiogenic gene can cause a switch from tumour dormancy to progressive tumour growth. Thirdly, tumour growth can be inhibited and sometimes tumour regression can be obtained just by attacking the vascular compartment with specific angiogenesis inhibitors. These three examples of preclinical experiments and many others have led to the conclusion that, in general, tumours are angiogenesis-dependent. Supported by disappointing clinical results, the angiogenesis dependency of tumours has been questioned, mainly because of the immaturity and the presumed lack of a functional blood supply (oxygen delivery and discarding of waste products) from a newly formed tumour vasculature. However, human tumours are highly heterogeneous in vascular architecture, differentiation, and functional blood supply. Vascular immaturity is a natural consequence of a genetically based unlimited expansion of tumour cells, compared to the well-regulated growth of different organs during embryonic development, for example. Unlimited tumour expansion and therefore the continuous stimulation of vessel outgrowth prevent endothelial cells from generating a mature vasculature, but instead continuously stimulate them to expand the vascular compartment of the growing tumour. In this review, the translation of angiogenesis inhibitors as a treatment for cancer from preclinical experiments to the clinic is evaluated. The preclinical evidence that tumours are angiogenesis-dependent is summarized and explanations are put forward for why the clinical results so far are not as exciting as was expected from preclinical studies. Reviewing the translation, one may conclude that human tumours are heterogeneous in their vascular architecture and function and that tumour-induced angiogenesis in humans is a more complex (multifactorially regulated) process compared with angiogenesis in preclinical cancer models. Copyright 2003 John Wiley & Sons, Ltd.