Passive microfluidic methods for sampling hormone secretions from primary islets and adipocytes

Abstract

We have developed an 8-channel passive microfluidic device that is capable of sampling both primary islet and adipocytes. Although there are microdevices already published in the literature; they require bulky external equipment and expertise to operate. In order to simplify operation of the device, fluidic resistance of the microchannels was used to control the flow rates through the device. In this way, a vacuum source, tubing, and a vacuum manifold are the only external components needed for operation. We thoroughly characterized the flow rate thorough the device by the meniscus tracking method, and determined at the pre-determined flow rate of 40 µL/h we would need a vacuum pressure of -7.7 kPa. The disadvantage to using PDMS as the channeled substrate is due to inherent surface characteristics; proteins tend to adsorb to the surface. A method was developed to prevent protein adsorption by treating the chip 1 h before islet sampling with imaging media that contained 0.1 % BSA. In order to limit the dead volume and prevent gradient formation of stimulants to the islets, a reservoir was designed to include all 8-channels of the device. Using different salt concentrations to mimic the glucose concentration change we measured the amount of time it took for the solution to reach the islet. With our new reservoir a total of 15 ± 2 s is the amount of time the islet will experience the glucose change, and since we are measuring the secretion for 1 h this time is negligible. We then loaded 8 islets on the device, and sampled secretions for 1 h, and compared these results to the standard bulk method. The two methods correlated well with each other (3 mM p < 0.8, 11 mM p < 0.65). Confocal images of the islets were obtained and volumes calculated. The volume of the islets was then compared to insulin secretion, and little to no correlation was found. The same 8-channel device was then adapted to sample primary adipocytes. By changing the reservoir design to encompass only a single channel, and including a moat region that sits above the entrance of the channel we are able to load and sample primary adipocytes using microfluidics. The adipocytes were stimulated with insulin and niacin, and the secretion of adiponectin was significantly higher than the un-treated adipocytes, for both treatments. To our knowledge, this is the first time primary adipocytes have been sampled and adiponectin quantified using microfluidics. By using conductivity solutions, a standard voltage meter, and voltage source we can directly measure the electrical resistance of a microfluidic channel, and then fluidic resistance can be calculated via a modified equation. Our new conductivity-based method was validated using 9 different resistors that had fluidic resistances varying between 40 – 600 kPa s mm^-3. The new conductivity-based method was compared, to the standard meniscus tracking method, and the two methods match well with one another. Also, there seems to be less error associated with the conductivity-based method than with the meniscus tracking method.