Poster P5 - Controlling Concentration Dynamics Using a Microfluidic Gradient Generator

Introduction: Concentration gradients are a key biophysical signal in many cell types and tissues, such as neurons, the intervertebral disc, and epithelial cells [1-3]. These gradients can vary both spatially and temporally in the body, directing the responses of cells and tissues based on the length or breadth of the response. While they are a key component of physiology, platforms that allow control of concentration gradients on the micro-scale are limited. Microfluidic devices offer an excellent opportunity to study cellular response to osmotic/concentration dynamics under highly controllable gradients due to the high control of solution and flow characteristics these devices afford. Here we present the characterization of a microfluidics-based device for the control of both temporal and spatial concentration gradients.

Materials and

Methods: PDMS microfluidic devices (Fig. 1A) were produced using soft lithography on photo-etched silicon wafers based off a previous design. 1mm Ø holes were punched at the outlet and 3 inlets of the device using a biopsy punch, then plasma bonded to a #1 glass coverslip. For testing, devices were first “primed” with deionized water to remove all air, and then 30% red food dye (McCormick) solution was flowed through by two Hamilton Syringe Pumps through 0.015in Tygon tubing attached to custom-made 90o needle connectors. To allow prediction of dye concentration from images, as described by Beer’s law, a concentration/absorbance curve was initially determined.

Concentration Kinetics Characterization:
Dye was flowed into inlet 3 only with water flowing into inlets 1&2 and run continuously for up to 2 hours. Brightfield images were collected every 30 seconds at each cell chamber (A-G) and the color intensity normalized to final timepoint in chamber G. The intensity over time for each chamber was fit to an exponential function in MATLAB (Fig 1F, left) quantifying the time constant for each chamber. Each experiment was repeated at least three times.

Steady-state Gradient Characterization:
Dye was flowed in multiple inlet orientations at multiple flowrates (0.5,1, and 5µL/min) and allowed to equilibrate for up to 1hr depending on the flowrate. At equilibrium image stitches of the gradient region were collected. For validation, a 2D digital copy of the microfluidic gradient generator was created within COMSOL and used to predict theoretical sodium gradient that would develop under various flow conditions.

Results, Conclusions, and Discussions:

Results: Beer’s Law correction showed excellent linear agreement (R2 = 0.9929, Data not Shown), and informed that a significant change in absorbance would be seen using a maximum concentration of 30% red dye by volume. Steady State experiment results recapitulated the behavior of modeled sodium concentration gradients through visual image comparison (Fig. 1B. Comparison of the two profiles shows excellent agreement with the sodium digital phantom and the steady state experiment (R2 = 0.918, RMSE = 0.118, Fig. 1C). This level of agreement between experiment and digital phantom did not vary based on configuration, as the 0-1-0 experiment also showed good visual and quantitative agreement with experiment (R2 = 0.925, RMSE = 0.082, Data not Shown). Time dependent experiment results showed behavior exponentially approaching an equilibrium concentration over time, with time constants dependent on inlet flowrates and chamber (Fig. 1D-F). 0.5uL/min experiments took the longest to reach equilibrium (τ = ~1hrs), while 5uL/min experiments took the shortest (τ = ~5mins). Results were unable to be obtained for chambers A and B, as in the 0-0-1 configuration (Mirror of 1-0-0) only water flows into those chambers.

Discussion: Results from the steady state experiments show great agreement with COMSOL digital phantom simulations of sodium chloride ion gradients, demonstrating the device’s capability to generate spatially varying concentration gradients. These results were independent of the configuration tested, demonstrating the model’s robustness in prediction. Time dependent experimental results also demonstrate the device’s ability to control concentration gradients through time with a predictable profile and rate of change. Calculated time constants for each of the flow rates match expected results, with the slowest changes being demonstrated in 0.5uL/min, and the fastest in 5uL/min.

Conclusions: This work introduces and demonstrates a novel, easy to manufacture, microfluidic platform to study concentration gradients varying both in time and in space. This study lays the groundwork for future studies focusing on cellular response to concentration dynamics and can be used as a tool to characterize the behavior of nanoscale ion sensors, such as DNA origami, to aid studies of mechanobiology.

Acknowledgements (Optional):

References: [1] Moore, et. al, Tissue Eng., 2006 [2] Bibby et. al., Spine, 2005 [3] Scherber et. al., Int. Bio, 2012
The authors would like to thank the Biomedical Engineering Department at The Ohio State University and NSF for their support.