Highly efficient cell-microbead encapsulation using dielectrophoresis-assisted dual-nanowell array

Abstract

Recent advancements in micro/nanofabrication techniques have led to the development of portable devices for high-throughput single-cell analysis through the isolation of individual target cells, which are then paired with functionalized microbeads. Compared with commercially available benchtop instruments, portable microfluidic devices can be more widely and cost-effectively adopted in single-cell transcriptome and proteome analysis. The sample utilization and cell pairing rate (∼33%) of current stochastic-based cell-bead pairing approaches are fundamentally limited by Poisson statistics. Despite versatile technologies having been proposed to reduce randomness during the cell-bead pairing process in order to statistically beat the Poisson limit, improvement of the overall pairing rate of a single cell to a single bead is typically based on increased operational complexity and extra instability. In this article, we present a dielectrophoresis (DEP)-assisted dual-nanowell array (ddNA) device, which employs an innovative microstructure design and operating process that decouples the bead- and cell-loading processes. Our ddNA design contains thousands of subnanoliter microwell pairs specifically tailored to fit both beads and cells. Interdigitated electrodes (IDEs) are placed below the microwell structure to introduce a DEP force on cells, yielding high single-cell capture and pairing rates. Experimental results with human embryonic kidney cells confirmed the suitability and reproducibility of our design. We achieved a single-bead capture rate of >97% and a cell-bead pairing rate of >75%. We anticipate that our device will enhance the application of single-cell analysis in practical clinical use and academic research.

Keywords: co-encapsulation; dielectrophoresis; hydrophilic beads; microfluidics; single-cell capture.

Fig. 1.

Design and configuration of the ddNA microfluidic device a) assembled ddNA microfluidic device. b) Microscope image of gold electrodes and the dual-nanowell array. The full chip contains 3,000 dual-nanowell structures separated into 10 individually controlled sectors. c) Microscopic close-up image of the nanowells and the nonsymmetric interdigitated electrodes. d) Cross-sectional schematic of the assembled device (not to scale). e) Top view and 3D schematic of the single ddNA unit. Single cells and single hydrophilic beads are captured in an interconnected dual-well structure.

Fig. 2.

A schematic illustration of the operational procedures of the ddNA device to realize the cell–bead co-encapsulation task. a, b) Hydrophilic beads are loaded into the device to preoccupy the larger trapping wells, and cells then flow into the channel with an AC power source. c) Cells are actively captured in smaller wells by DEP force. d) Fluorinated oil is introduced into the channel to seal the microwell array, and a freeze-thaw cycle lyses the cells in an encapsulated microreaction chamber.

Fig. 3.

Evaluation of bead-loading occupancy and trapping rate with different designs and flow rates. a) Efficient bead loading achieves >97% occupancy through perfusing, trapping, and washing steps. b) Relationship between bead-trapping efficiency and washing flow rate for different well depths. Each data point shows the mean and standard deviation of the percentage of trapped beads at randomly chosen locations in >200 ddNA units in three independent experiments performed on the same chip. c) Demonstration of beads trapped in 20 and 40 μm depths under different flow rates.

Fig. 4.

Evaluation of DEP-assisted cell pairing, pairing rate, and the spatial magnitude distribution of DEP force. a) Cell–bead co-capture in ddNA units. b) Fluorescent image of the co-captured cell–bead pairs. Most of the beads were paired with a single cell, enabling higher bead utilization compared with other microfluidic systems. c) Relationship between the flow rate and single-cell pairing rate in a continuous flow cell-loading manner. Each bar shows the mean and standard deviation of the three experiments repeated on different sectors of the same chip. More than half of the ddNA units chosen at random locations were counted. d) Numerical simulation of the magnitude of $\mathrm{\nabla }|\mathit{E}{|}^2$ vector at the z = 10 μm plane above a ddNA unit.

Fig. 5.

Cell lysis inside enclosed ddNA units. a) Fluorescent image of captured cells in ddNA units before and after (7× exposure) a 10 min freeze-thaw cycle. The dual-nanowell structures have been sealed by the fluorinated oil. b) Fluorescent image of captured cells in the control experiment before and after 15 min at room temperature. A change in cell viability was not observed. c) Change in radial fluorescence intensity due to cell lysis. Fluorescence intensity values were obtained along the solid line and dashed line in image a).