A Microfluidic Hanging-Drop-Based Islet Perifusion System for Studying Glucose-Stimulated Insulin Secretion From Multiple Individual Pancreatic Islets

Abstract

Islet perifusion systems can be used to monitor the highly dynamic insulin release of pancreatic islets in glucose-stimulated insulin secretion (GSIS) assays. Here, we present a new generation of the microfluidic hanging-drop-based islet perifusion platform that was developed to study the alterations in insulin secretion dynamics from single pancreatic islet microtissues at high temporal resolution. The platform was completely redesigned to increase experimental throughput and to reduce operational complexity. The experimental throughput was increased fourfold by implementing a network of interconnected hanging drops, which allows for performing GSIS assays with four individual islet microtissues in parallel with a sampling interval of 30 s. We introduced a self-regulating drop-height mechanism that enables continuous flow and maintains a constant liquid volume in the chip, which enables simple and robust operation. Upon glucose stimulation, reproducible biphasic insulin release was simultaneously observed from all islets in the system. The measured insulin concentrations showed low sample-to-sample variation as a consequence of precise liquid handling with stable drop volumes, equal flow rates in the channels, and accurately controlled sampling volumes in all four drops. The presented device will be a valuable tool in islet and diabetes research for studying dynamic insulin secretion from individual pancreatic islets.

Keywords: glucose-stimulated insulin secretion (GSIS); hanging drops; microfluidics; organ on chip; pancreatic islets; perifusion systems.

FIGURE 1

Theory and concept of the automatic hanging-drop-size adjustment in the microfluidic chip. (A) The inflow (Q_in) is split into two flows, one toward the islet drop (right) and the other toward the control drop (left). There is a continuous flow of liquid from the inlet toward the islet drop due to active sampling from the sampling outlet at the rate of Q_{out sampling}. An irregular flow is observed from the inlet toward the control drop due to the alternating withdrawal of liquid and air from the needle-type valve at a rate of Q_{out control}. (B) Self-regulation of drop heights between two interconnected hanging drops. Liquid is constantly added into the system at a rate of Q_in and withdrawn through the sampling outlet at a rate Q_{out S} and through a needle-type valve at a rate of Q_{out C}. (i) Two hanging drops in equilibrium with identical Laplace radius and pressure. The inflow is evenly distributed to the two drops. As the tip of the needle-type valve is immersed in the drop, liquid removal through the valve will occur. As a consequence, (ii) the size and the Laplace pressure of the two drops becomes different with drop radius and pressure being inversely correlated. Due to the pressure difference, there is an increased flow toward the control drop (left) with the lower pressure. As soon as the tip of the needle-type valve is exposed to air, only air is withdrawn. The control drop remains stable at the height defined by the valve needle length. (iii) In a next step the two drops reach equal Laplace radius and pressure through equilibration through the liquid phase. The system is now stable, and the drop height in control and tissue drop is maintained constant.

FIGURE 2

The microfluidic hanging-drop perifusion system. (A) Chip layout and dimensions. The fluidic structures (light blue) have a depth of 500 μm, except for the common inlet that has a recess depth of 1 mm. The hydrophobic rim structures (white) define the fluidic channels and hanging drops. They have a height of 250 μm measured from the chip surface. Highlighted channel sections in dark blue and orange were considered for the calculation of the hydraulic resistance in the channels. The channels between common inlet and control drop and between common inlet and islet drops were designed to have the same hydraulic resistance. (1) Cross-sectional view of the islet drop with the islet microtissue at the bottom of the hanging drop. (B) Top view of the assembled chip with (i) four parallel outlet tubes, (ii) one common inlet tube, and (iii) a NanoPort assembly with a needle-type valve inserted at the center. Scale bar: 10 mm. (C) Side views of the chip with hanging drops visualized with red dye. All hanging drops have equal sizes and shapes. Scale bar: 5 mm.

FIGURE 3

Experimental setup. Two syringes, filled with different media, are connected through a splitter to the common inlet of the chip. Inflow of medium is actuated by using a neMESYS syringe pump. One peRYSIS peristaltic pump was used to actively withdraw liquid from the system at a constant equal flow rate through the 4 sampling outlets and at a higher flow rate (larger tubing diameter) through the control drop needle. Samples were collected into a 384-well plate using the rotAXYS sampling arm (Supplementary Figure 1B) for further downstream analysis. The medium withdrawn from the control drop went to waste.

FIGURE 4

Fluidic system characterization. (A) Drop height fluctuations of an islet drop over a 73 min perifusion sequence at 37°C. The red dotted line indicates the mean drop height. A close-up is shown for a 5-min duration. (B) Temporal characteristics of drop size fluctuations and of withdrawal of air-liquid fractions through the needle-type valves. Data are mean values ± standard deviation of measurements from eight independent experiments. An average of four to six measurements (5–60 min time window) were taken per experiment. (C) Concentration of NaFl, sampled from each hanging drop during a 44 min perifusion sequence. 5 μM NaFl was continuously injected at 500 nL min^{– 1} directly into the hanging drops through a fine needle, while 7.5 μL samples were collected in a well plate at 30 s intervals.

FIGURE 5

Modeling results of the insulin concentration profiles at the chip outlet after a sharp 1-s-long insulin secretion burst from an islet microtissue (black trace) for different parameters. (A) Different drop heights with an islet positioned in the center of the drop at a perifusion rate of 15 μL min^–1. (B) Insulin concentration measured in the well plate for a sampling rate of 30 s. (C) Modeling results for constant and continuous insulin secretion at different drop heights with an islet positioned in the center of the drop at a perifusion rate of 15 μL min^–1. (D) Different perifusion rates in a 0.6 mm drop with an islet positioned in the center of the drop.

FIGURE 6

High-resolution microfluidic FlowGSIS measurements of four single islet microtissues in parallel. A change from low (2.8 × 10^–3M) to high (16.7 × 10^–3M) glucose concentrations stimulated similar biphasic insulin secretion responses in all four islets. Samples were continuously taken every 30 s between 0 and 45 min, and every 60 s between 45 and 66 min.