Vascularization is a prerequisite for tumour growth and metastasis. Therefore anti-angiogenic therapy is one of the most promising strategies to defend against cancer. Consequently a large number of anti-angiogenic agents is currently tested in animal experiments that include the rabbit corneal assay, the mouse cheek pouch assay, and the mouse and rabbit cranial and skin window. Anti-angiogenic substances, such as thalidomide, are nowadays routinely used in several cancer therapies, for example the treatment of multiple myeloma. Preclinical studies of such substances require a large number of animal experiments. Here we introduce an in vitro model for anti-angiogenic screening based on co-culturing of embryonic stem cell-derived embryoid bodies and multicellular tumour spheroids, in order to reduce the number of animals needed for in vivo testing. The embryoid body derived from embryonic stem cells mimics early stages of embryogenesis, facilitating research on various differentiation processes of all three germ layers. This system combined with multicellular tumor spheroids simulates interactions between healthy and pathological tissue, thus mimicking the process of tumour- induced angiogenesis. We observed that blood vessels produced in co-culture are denser than in embryoid bodies grown alone, and also show directed growth of vessels towards the tumour. The anti-angiogenic and dose-dependent effect of three substances (thalidomide and the tyrosine kinase inhibitors SU5614 and ZM323881) is shown by immune-cytochemistry and flow cytometry, based on staining for PECAM-1 (CD31) and VE-Cadherin (CD144). FACS analyses indicate that the number of CD31+ cells is not significantly changed, but the degree of vascularization is reduced upon treatment with anti-angiogenic agents. It is also shown that tyrosine kinase inhibitors have anti-inflammatory properties besides being anti-angiogenic, which is visualised by staining for leucocyte specific markers CD45, CD68 and neutrophil antigen. Preventing inflammation is of importance, since it often occurs next to the tumour, thus inducing a pro-angiogenic micro-environment.

In summary we demonstrate that confrontation cultures of embryoid bodies and multicellular tumor spheroids are suitable to replace or reduce the number of animal in vivo studies, e.g. the rabbit cornea assay, the hamster cheek pouch assay, the cranial window assay or the skin chamber assay. Our novel in vitro confrontation culture model can be used for screenings of various substances as well as for examinations of signalling pathways involved in host tissue-tumor interaction.