Optimization of Microfluidic Cell Encapsulation

Islet cell transplantation is a type of cell therapy that provides a promising treatment avenue for type-1 diabetic patients. As a minimally invasive procedure, islet transplantation can achieve patient insulin independence. However, the body’s immune response can lead to patients reverting to insulin dependence. One way to circumvent this is to encapsulate the islet cells by using alginate, a biomaterial, to form microdroplets which can then be introduced to ionic cross-linking agents (e.g., Ca²⁺). Here a surface reaction called chelation occurs between the alginate and the divalent cation, forming a coating. Small molecules can cross the coating, like insulin, oxygen, and nutrients. However, larger entities, like immune cells and antibodies cannot cross the barrier.

One method to produce these microcapsules is via a microfluidic device. These devices house microchannels that allow an inlet solution of alginate to be cut into droplets by a secondary, immiscible fluid. The size and surface area of the alginate-based hydrogel are crucial to the kinetic control of insulin release. This technique allows for more precise manipulation of droplet formation, enhances quality control over traditional encapsulation methods, and the ability to better tune the consistency, size, and shape of droplets.

Testing every possible device geometry is not feasible. However, if a dynamic computational model of droplet generation could be built to simulate the microfluidic process, one could then design, test, and improve the microfluidic devices more efficiently.

To address this question, three unique microfluidic devices were first designed using AutoCAD 2019 (Autodesk, Inc.). Then, the designs were processed in the computational fluid dynamics software, ANSYS Fluent 2020 (Ansys, Inc.), and fabricated. The flow ratios between the carrier fluid and the alginate flow were adjusted in a designed experiment, and the effects on size, shape, and droplet formations were used as comparison metrics between the real and simulated behavior.

In this project, different designs were compared to their respective simulated behavior to test whether equivalent flow rate ratios yielded similar flow behaviors at the junctions and similar droplet consistency and shape. Our preliminary results show that actual and simulated flow settings required rates of different magnitudes, and droplet diameters were dissimilar. Analysis of the experimental methodology suggests that fine-tuning the CFD software promises to accurately mimic real microfluidic device behavior with a more robust knowledge of the software and more accurate physical parameters (e.g., surface / interfacial tensions).