Development and Optimization of Microfluidic Devices for the Study of Free Fatty Acids

Free fatty acids (FFAs) are signaling molecules secreted by adipocytes, or fat cells. Individual FFAs differ in their ability to stimulate or attenuate insulin secretion depending on the chain length and degree of unsaturation of the FFA. Due to these varying effects of individual FFA on insulin release and the relationship of obesity and type 2 diabetes, it is important to temporally resolve the secretion of individual FFA from adipocytes. While the selective re-incorporation and secretion of FFAs have been extensively studied, few have monitored the temporal secretory profiles of individual FFAs to clarity their biological effects. One of the major obstacles in monitoring temporal secretory profiles of FFAs is the time- and labor-intensive sample preparation for analysis by gas chromatography-mass spectrometry (GC-MS). Conventional preparation procedures are often inefficient since multiple extraction steps are required followed by derivatization to fatty acid methyl esters (FAMEs). The total time can require up to 48 hours before samples can be quantified by a GC-MS method. These requirements hinder sample throughput and are often prohibitive to the detection of acute changes in FFA secretions. Microfluidics can address these drawbacks by reducing the amount of reagents consumed and automating the procedures by integrating sample preparation steps. This dissertation describes the development and optimization of an integrated microfluidic device and associated analytical methods that could be used to monitor individual FFAs to assess what role they might have in cellular communication between fat tissue and the pancreas. The first module of the integrated device automated the derivatization of FFA to FAME via acid catalyzed esterification using methanolic-HCl reagent. A statistical multivariate optimization protocol called Design of Experiment (DOE) was used to aid in the method development by determining the combination of derivatization time (Tder) and ratio of methanolic-HCl:FFA (Rder) that maximized the conversion of two model FFAs to their methyl ester forms. Optimal derivatization conditions for the two model FFAs was Tder = 0.8 min and Rder = 4.9 with the resulting total sample preparation time of 5 min. This combination of Tder and Rder was used to derivatize 12 FFAs with a range of derivatization efficiencies from 18% to 93%. As compared to a conventional macroscale derivatization of FFA, the derivatization module decreased the volume of methanolic-HCl and FFA by 20- and 1300-fold, respectively. The second module of the microfluidic device purified the derivatized sample by liquid-liquid extraction of FAME from the derivatization fluid into a hexane phase for GC-MS analysis. The derivatization module was coupled to the extraction module in a continuous-flow format. Phase contact with the hexane flow in stratified and segmented flow formats were qualitatively investigated as methods to extract FAME. The flow regimes in which both flow formats can be generated were explored and applied to generate parallel fluid flow or alternating aqueous and organic segments. Further work is needed to separate the two phases in both formats before the integrated device can be applied to study the temporal secretion of individual FFAs from adipocytes. A second goal of this research was to design and numerically optimize a microfluidic perfusion chamber capable of housing a large number of pancreatic islets of Langerhans for the application of glucose concentration in sinusoidal waveforms. A DOE tool called response surface methodology (RSM) was used to determine the optimum perfusion chamber size and operating flow rate that would deliver homogeneous glucose waveforms to 40 islets of Langerhans with minimal shear stress. 3D computational models were used to calculate the homogeneity of glucose concentration and magnitude of shear rate within the chamber. The optimal perfusion chamber design with chamber length of 1.84 mm and a flow rate of 10 μL min-1 was validated to show that the 40 islets experienced shear rate ≤ 50 s-1 and a 0.13% variability in the spatial distribution of glucose within the chamber. The optimized design will permit the sampling of secretory products in a format that can be easily incorporated in an online analysis to realize an integrated system for the quantification of multi-analyte secretory profiles from islets. Separately, the devices described in this dissertation allow for the assessment of adipocytes and islet functions. When integrated in a single device, it will permit investigation of how temporal secretion patterns for individual FFAs from adipocytes might alter the signaling processes within the islets of Langerhans that lead to type 2 diabetes.