Creating capillary networks within human engineered tissues: impact of adipocytes and their secretory products

The development of tissue-engineered substitutes of substantial volume is closely associated with the need to ensure rapid vascularization upon grafting. Strategies promoting angiogenesis include the in vitro formation of capillary-like networks within engineered substitutes. We generated both connective and adipose tissues based on a cell sheet technology using human adipose-derived stromal cells. This study evaluates the morphology and extent of the capillary networks that developed upon seeding of human microvascular endothelial cells during tissue production. We posited that adipocyte presence/secretory products could modulate the resulting capillary network when compared to connective substitutes. Analyses including confocal imaging of CD31-labeled capillary-like networks indicated slight differences in their morphological appearance. However, the total volume occupied by the networks as well as the frequency distribution of the structure's volumes were similar between connective and adipose tissues. The average diameter of the capillary structures tended to be 20% higher in reconstructed adipose tissues. Quantification of pro-angiogenic molecules in conditioned media showed greater amounts of leptin (15×), angiopoietin-1 (3.4×) and HGF (1.7×) secreted from adipose than connective tissues at the time of endothelial cell seeding. However, this difference was attenuated during the following coculture period in endothelial cell-containing media, correlating with the minor differences noted between the networks. Taken together, we developed a protocol allowing reconstruction of both connective and adipose tissues featuring well-developed capillary networks in vitro. We performed a detailed characterization of the network architecture within engineered tissues that is relevant for graft assessment before implantation as well as for in vitro screening of angiogenic modulators using three-dimensional models.