Resumen de Study of the diffusive properties of extracellular matrix gels used in 3D cultures
Roland Galgoczy
The extracellular matrix (ECM) hinders diffusion in both tissues and hydrogels used in 3D cultures. However, how geometric and non-geometric ECM properties contribute to diffusion hindrance remains poorly understood. To address this question, the effective diffusivity D of FITC-dextrans with molecular weight (Mr, 4-70 kDa) spanning the physiological range of signaling factors was measured in a panel of acellular ECM gels with an optical macroscopic assay. Gels included reconstituted basement membrane/Matrigel, fibrin and type I collagen, and exhibited an average pore size much larger than any dextran size. Unexpectedly, a decay of D with Mr following a power-law with an exponent that matched that predicted by the Stokes-Einstein law in all gels (D ~ Mr-1/3, r2=0.99) was observed, revealing that hindered diffusion is dominated by nongeometric viscous factors. This law predicted that Matrigel and fibrin exhibited similar viscosities, which was confirmed with microrheology measurements by atomic force microscopy. Moreover, gels with the lowest D exhibited diffusion hindrance comparable to the extreme physiologic hindrance of brain tissue, which has a typical pore size much smaller than ECM gels. In contrast, diffusion hindrance in sparse (/=3 mg/ml) are suitable tissue surrogates in terms of macromolecular diffusion. Additionally, it is unknown to what extent current 3D culture protocols provide physiologic oxygen tension conditions. To address this limitation, oxygen tension was measured within the acellular or cellularized ECM gels with A549 cells, and analyzed in terms of oxygen diffusion and consumption. Oxygen diffusivity in acellular gels was up to 40% smaller than that of water, and the lower values were observed in the denser gels. In 3D cultures, physiologic oxygen tension was achieved after 2 days in dense (>/=3 mg/ml) but not sparse gels, revealing that the latter gels are not suitable tissue surrogates in terms of oxygen distribution. In dense gels, a dominant effect of ECM composition over density in oxygen consumption as observed. All diffusion and consumption data were used in a simple model to estimate ranges for gel thickness, seeding density and time-window that may support physiologic oxygen tension. Thus, critical variables for oxygen tension in ECM gels were identified, and a model to assess initial values of these variables was introduced, which may short-cut the optimization step of 3D culture studies.
