Reversibly sealed multilayer microfluidic device for integrated cell perfusion and on-line chemical analysis of cultured adipocyte secretions

Abstract

A three-layer microfluidic device was developed that combined perfusion of cultured cells with on-line chemical analysis for near real-time monitoring of cellular secretions. Two layers were reversibly sealed to form a cell chamber that allowed cells grown on coverslips to be loaded directly into the chip. The outlet of the chamber was in fluidic contact with a third layer that was permanently bonded. Perfusate from the cell chamber flowed into this third layer where a fluorescence enzyme assay for non-esterified fatty acid (NEFA) was performed on-line. The device was used to monitor efflux of NEFAs from approximately 6,200 cultured adipocytes with 83 s temporal resolution. Perfusion of murine 3T3-L1 cultured adipocytes resulted in an average basal concentration of 24.2 +/- 2.4 microM NEFA (SEM, n = 6) detected in the effluent corresponding to 3.31 x 10(-5) nmol cell(-1) min(-1). Upon pharmacological treatment with a beta-adrenergic agonist to stimulate lipolysis, a 6.9 +/- 0.7-fold (SEM, n = 6) sustained increase in NEFA secretion was observed. This multilayer device provides a versatile platform that could be adapted for use with other cell types to study corresponding cellular secretions in near real-time.

Fig. 1

Chip Design (a) The multilayer device was comprised of three separately etched glass wafers that integrated a cell perfusion chamber and fluidic channels for on-line mixing of the fluorescence-based enzyme assay. (b) A side view of the cell chamber depicts the bonded and reversibly-sealed portions of the device. (c) An in-house built aluminum compression frame functioned to compress the three glass wafers together enabling reversible sealing as well as to serve as a microscope stage holder.

Fig. 2

Characterization of Fluid Dynamics of Multilayer Device (a) A diagram of the characterization experiment and the three detection points are shown. The change in fluorescence resulting from a step change was observed along the flow path of the chip at the end of the inlet capillary to the chip (point 1), prior to the mixing point of the enzyme assay (point 2), and at the detection spot (point 3). (b) Plot of the normalized resorufin step changes at different detection points. Rise times shown were calculated from the plots.

Fig. 3

Fluidic Modeling of NEFA Flux (a) Side view image of channel showing predicted NEFA concentration distribution within cell chamber at different times following a 6-fold step increase in NEFA efflux. (b) Simulations of the 6-fold step change. Model predicts a 56 s 10 – 90% rise time following step change. Arrow indicates time when the step change started (t = 500 s).

Fig. 4

On-Line NEFA Enzyme Assay Reaction and Calibration (a) The scheme of the fluorescence-based enzyme assay for NEFAs used for on-line mixing and detection is shown. Acyl-CoA synthetase (ACS), acyl-CoA oxidase (ACOD), peroxidase (HRP), and Amplex UltraRed were mixed on-chip with NEFAs (either from standards or effluent) to form a fluorescent product similar to resorufin. The R group represents a fatty acid tail of 3 to 17 carbon lengths (6 to 20 total carbons in length). (b) An example step-change calibration using the multilayer device is depicted. NEFA standards were flowed through the device ranging between 0 μM – 250 μM. (c) Calibration curve generated from the data in Figure 4a.

Fig. 5

On-line NEFA Secretion Monitoring from Cultured Adipocytes (a and b) Representative traces of NEFA release from differentiated adipocytes and response from treatment with 20 μM isoproterenol (indicated by the bar) are shown. (c) A plot of the averaged normalized release depicts a 6.9 ± 0.7-fold (SEM, n = 6) increase in release of NEFAs upon isoproterenol treatment. The grey lines above and below the average response are ± SEM. (d) A control experiment was performed that monitored NEFA release in the presence and absence of enzyme reagents. Basal and isoproterenol-treated NEFA values were obtained once the enzyme reagents were added back to the system. The insert depicts the values detected during basal and isoproterenol treatment without enzyme reagents present.