A prerequisite for successful tissue engineering is the existence of a functional microvascular network. We hypothesized that such networks can be created and quantified in an in vitro setting by co-culturing endothelial cells (ECs) with tissue-specific 'bystander cells' in 3-D gel matrices. To test this hypothesis we adapted a previously described in vitro microcarrier-based angiogenesis assay (V. Nehls and D. Drenckhahn, 1995, Microvasc Res 50: 311-322). On optimizing this assay, we noted that the initial EC-microcarrier coverage depended on EC type and seeding technique employed to coat the microcarrier beads with the ECs. A confluent EC monolayer on the microcarrier surfaces formed only when bovine aortic endothelial cells (BAECs) were admixed to the beads under gentle agitation on an orbital shaker. After embedding BAEC-covered microcarrier beads into a sandwich-like arrangement of collagen or fibrin gels, we assessed cellular outgrowth at different serum concentrations in terms of migration distance and sprout formation. Quantifiable sprout formation was highest at 1% fetal bovine serum (FBS) in collagen matrices and at 0.1% FBS in fibrin matrices. At higher serum concentration, excess cell migration and formation of clusters prevented quantitative analysis of sprouting. Following the fine-tuning of this angiogenesis assay, we co-cultured BAECs with adipose tissue-derived fibroblasts (FBs) and vascular smooth muscle cells (SMCs). While FBs were able to increase the average migration distance of BAECs in both matrices, SMCs enhanced BAEC migration in fibrin, but not in collagen gels. By contrast, the number of newly formed sprouts in fibrin gels was increased by both cell types. We conclude that in this model bystander cells enhance EC network formation in a matrix-dependent manner. Additionally, these results stress the importance of carefully selecting the experimental parameters of a given in vitro angiogenesis model.