In this case, the different stages of angiogenesis can be a target for drug administration: cell migration, proliferation and differentiation. Since the differentiation stage of angiogenesis is the final and most important step during which capillary-like tubules are formed, the assays simulating this stage are extensively used to determine the efficacy of anti-angiogenic drugs. Currently used differentiation assays involve the plating of endothelial cells into or onto a gel matrix to investigate the formation of tubule-like structures after 4 to 24 h. The Matrigel assay is a widely used differentiation assay, where a protein mixture derived from mouse Engelbreth-HolmSwarm sarcoma is used. After 12 h of endothelial cell seeding on the Matrigel, the cells start to form cord-like structures, although in many cases, the seeded cells often clump into cell aggregates. Additionally, there is significant debate as to whether these cordlike structures actually contain patent lumina or not. Furthermore, non-endothelial cells like fibroblasts and glioblastoma cells have also been shown to form cord-like structures on Matrigel, and therefore a cautious interpretation of the results in this system is required. A further limitation of these gel assays is their inability to mimic the complete in vivo situation, since the surrounding cells near an angiogenic site are not taken into consideration. In order to test angiostatic drugs efficiently, definitive in vitro assays are necessary to reliably estimate the anti-angiogenic potential of the drugs. 2-D assays should be fast, inexpensive and easy to set up. Furthermore, the assays should optimally imitate the in vivo situation. Most translatable assays would include supporting cells, extracellular matrix and/or basement membrane. Circulating blood would further optimize an in vitro assay. After an estimation of applicability via the 2-D assay, an in vivo assay could then be considered for further analysis. Co-culture models could imitate the in vivo situation in a more reliable manner than Matrigel or other ECM models, since supportive cells adjacent to the angiogenic site in the body also have an influence on vessel development. Several co-culture models have been described so far, using HUVECs or human dermal microvascular endothelial cells as sproutbuilding cells, and different supportive cell types in 2D and 3D. These supportive cells develop an ECM and/or release growth factors that support angiogenesis. As supportive cells in co-culture models, fibroblasts from dermal human breast tissue, HUASMCs, pulmonary RAD001 artery smooth muscle cells, mesenchymal stem cells and fibroblasts from dermal superficial skin layer have been used. All of these supportive cells in co-culture with sprout-forming cells supported the development of capillarylike structures in 2D. These co-cultures provide a reliable estimation regarding the anti-angiogenic potential of a particular drug. Unfortunately, all supportive cells used in the aforementioned assays derive from human tissue cell sources with limited access.