Variable optimization for the formation of three-dimensional self-organized heart muscle

Cardiac tissue-engineering research is focused on the development of functional three-dimensional (3D) heart muscle in vitro. These models allow the detailed study of critical events in organogenesis, such as the establishment of cell–cell communication and construction and modification of the extracellular matrix. We have previously described a model for 3D heart muscle, termed cardioids, formed by the spontaneous delamination of a cohesive monolayer of primary cells in the absence of any synthetic scaffolding material. In an earlier publication, we have shown that, upon electrical stimulation, cardioids generate a twitch force in the range of 200–300 µN, generate a specific force (twitch force normalized to total cross-sectional area) of 2–4 kN/m2, and can be electrically paced at frequencies of up to 10 Hz without any notable fatigue. We have two objectives for the current study: model development and model optimization. Our model development efforts are focused on providing additional characterization of the cardioid model. In this study, we show for the first time that cardioids show a pattern of gene expression comparable to that of cells cultured in two dimensions on tissue culture plastic and normal mammalian heart muscle. Compared with primary cardiac cells cultured on tissue culture plastic, the expression of α-myosin heavy chain (MHC), β-MHC, SERCA2, and phospholamban was significantly higher in cardioids. Our second objective, model optimization, is focused on evaluating the effect of several cell culture variables on cardioid formation and function. Specifically, we looked at the effect of plating density (1.0–4.0 × 106 cells per cardioid), concentration of two adhesion proteins (laminin at 0.2–2.0 µg/cm2 and fibronectin at 1–10 µg/cm2), myocyte purity (using preplating times of 15 and 60 min), and ascorbic acid stimulation (1–100 µl/ml). For our optimization studies, we utilized twitch force in response to electrical stimulation as our endpoint metric. Based on these studies, we found that cardioids formed with a plating density in the range 3–4 × 106 cells per cardioid generated the maximum twitch force, whereas increasing the surface adhesion protein (using either laminin or fibronectin) and increasing the myocyte purity both resulted in a decrease in twitch force. In addition, increasing the ascorbic acid concentration resulted in an increase in the baseline force of cardioids, which was recorded in the absence of electrical stimulation. Based on the model development studies, we have shown that cardioids do indeed exhibit a gene expression pattern similar to normal mammalian heart muscle. This provides further validity for the cardioid model. Based on the model optimization studies, we have identified specific cell culture regimes which support cardioid formation and function. These results are specific to the cardioid model; however, they may be translated and applied to other tissue-engineering models. Collectively, the work described in this study provides insight into the formation of functional 3D heart muscle and the effect of several cell culture variables on tissue formation and function.