A Computational Model of a Microfluidic Device to Measure the Dynamics of Oxygen-Dependent ATP Release from Erythrocytes

摘要

Erythrocytes are proposed to be involved in blood flow regulation through both shear- and oxygen-dependent mechanisms for the release of adenosine triphosphate (ATP), a potent vasodilator. In a recent study, the dynamics of shear-dependent ATP release from erythrocytes was measured using a microfluidic device with a constriction in the channel to increase shear stress. The brief period of increased shear stress resulted in ATP release within 25 to 75 milliseconds downstream of the constriction. The long-term goal of our research is to apply a similar approach to determine the dynamics of oxygen-dependent ATP release. In the place of the constriction, an oxygen permeable membrane would be used to decrease the hemoglobin oxygen saturation of erythrocytes flowing through the channel. This paper describes the first stage in achieving that goal, the development of a computational model of the proposed experimental system to determine the feasibility of altering oxygen saturation rapidly enough to measure ATP release dynamics. The computational model was constructed based on hemodynamics, molecular transport of oxygen and ATP, kinetics of luciferin/luciferase reaction for reporting ATP concentrations, light absorption by hemoglobin, and sensor characteristics. A linear model of oxygen saturation-dependent ATP release with variable time delay was used in this study. The computational results demonstrate that a microfluidic device with a 100 µm deep channel will cause a rapid decrease in oxygen saturation over the oxygen permeable membrane that yields a measurable light intensity profile for a change in rate of ATP release from erythrocytes on a timescale as short as 25 milliseconds. The simulation also demonstrates that the complex dynamics of ATP release from erythrocytes combined with the consumption by luciferin/luciferase in a flowing system results in light intensity values that do not simply correlate with ATP concentrations. A computational model is required for proper interpretation of experimental data.