Microfluidic technology for multiple single-cell capture

Abstract

Microfluidic devices are widely used in single-cell capture and for pairing single cells or groups of cells for cell–cell interaction analysis; these advances have improved drug screening and cell signal transduction analysis. The complex in vivo environment involves interactions between two cells and among multiple cells of the same or different phenotypes. This study reviewed the core principles and performance of several microfluidic multiple- and single-cell capture methods, namely, the microwell, valve, trap, and droplet methods. The advantages and disadvantages of the methods were compared, and suggestions regarding their application to multiple-cell capture were provided. The results may serve as a reference for research on microfluidic multiple single-cell coculture technology.

I. INTRODUCTION

The responses of individual cells in cell–cell interactions vary over time. Their temporal behavior and phenotypes cannot be observed without a time lapse or continuous monitoring. Therefore, observing the phenotype of each cell through time-lapse microscopy is essential for analyzing cell–cell interactions. A cell's DNA, RNA, and protein expression change depending on the stage of the cell cycle.¹ This results in changes in cell morphology, which can increase or decrease the area of contact between cells and affect cell–cell interactions.² Cell–cell interactions are crucial to tissue development and physiological function.³ Cells communicate with neighboring cells through the secretion of signal molecules or through cell–cell contact in tissue. Cells of the same type may have different interaction mechanisms because of their heterogeneity. Therefore, investigating the interaction between cells is essential. Key physiological functions such as wound healing, cancer cell proliferation, metastasis, and stem cell development are related to cell–cell interaction.^4–6 The cell coculture method is widely used to study interactions between cells (Fig. 1).⁷ The method involves culturing two or more types of cells in the same device to analyze the behavior of or expression in cells. In laboratories, a conventional transwell plate, in which two types of cells are cocultured on either side of a permeable membrane, is often used for coculturing. Thus, transwell permeable membranes allow for the transmission of cell secretion signal molecules between cells, thus promoting cell–cell interaction.⁸ Cell communication is conducted through contact interaction (Fig. 1, direct coculture) and noncontact interaction (Fig. 1, indirect coculture). Noncontact interaction relies on the diffusion of the signal molecules secreted by a cell through extracellular fluid to the surface of another cell and their binding to the membrane proteins on the surface to transmit information. Contact interaction occurs through the transmission of signal molecules through gap junctions for communication.⁹ The transwell coculture method can only be used to analyze the population average formed at the end of contact and noncontact interaction; it cannot be used to study phenotype or genotype diversity within a cell population.¹⁰ Therefore, studies using the transwell coculture method have often ignored the specific expression of rare cell populations (such as cancer stem cells) because the population average often obscures the data of these specific values,¹⁰ causing inconsistency among research results (Fig. 2).¹¹ Cells that are particularly resistant to a drug cannot be identified during drug screening, which results in poor prognosis after treatment. The design of conventional transwell plates makes analyzing and counting individual heterogeneous cells difficult, thereby making the accurate analysis of cell–cell interactions challenging. This limitation makes identifying the complex mechanisms underlying cell–cell interaction difficult, especially in highly heterogeneous tumor cells.¹² The cocultivation of multiple cell types in an environment similar to a microenvironment in vivo would be required to identify the mechanism of cell interactions.¹³ However, the detection of cell subpopulations and analysis of single-cell behavior through transwell cocultivation systems are often inaccurate. With advances in microfabrication and microfluidic technology, complex and miniaturized microsystem devices have increasingly been used in biomedical research. These devices are mainly fabricated through soft lithography, etching, electron beam lithography, and 3D printing technology using biocompatible substrates such as polydimethylsiloxane, plastic, and glass. These small devices can be combined with pumps, valves, filters, and sensors to create a multifunctional lab-on-a-chip system.¹⁴ This lab-on-a-chip system can overcome the limitations of conventional experimental methods because of its precision temperature control, gas control, culture space geometry, nutrient supply control, culture environment control, high throughput, and automated culture capacity. Microfluidic systems have been widely used for cell sampling, cell capture, cell sorting, cell arrangement, drug screening, and multiparameter analysis.¹⁵ However, many bottlenecks remain in the development of coculture microfluidic chips used to capture multiple cells. This paper introduces a single-cell capture technology and describes the current method of pairing single cells or groups of single cells. A comparison of the efficacy and limitations of the various methods can provide insight for the future development of multiple single-cell capture technology. The results of this study may serve as reference for researchers interested in microfluidic multiple single-cell coculture technology.

FIG. 1.

Experimental coculture model. Reproduced with permission from Michelle et al., Front Bioeng. Biotechnol. 8, 911 (2020). Copyright 2020 Vis, Ito, and Hofmann.

FIG. 2.

Single-cell analysis revealing heterogeneity. Reproduced with permission from Fang et al., J. Hematol. Oncol. 10(1), 27 (2017). Copyright 2017 Springer Nature.

II. MULTIPLE- AND SINGLE-CELL CAPTURE METHODS

In in vivo environments, various types of cells are distributed in a 3D space, and these cells constantly interact with neighboring cells and transmit complex responses to each other. This dynamic communication is crucial to maintaining tissue function, regeneration, and repair.^5,17 However, tracking and analyzing the number of cell signals, the effect of distance on signal transmission, and signal interference between cells are difficult. Tracking the behavior of individual cells requires the integration of multiple methods. We compared the efficacy of four methods: the microwell, valve, trap,¹⁵ and droplet microfluidic capture methods. The efficiency of single-cell capture methods is often evaluated in terms of cell viability, capture efficiency, throughput, cultivation, ability to isolate cells directly, downstream analysis, number of cells required, and operating time (Table I). Cell damage must be prevented during capture because the shear stress of the fluid, physical damage, and environmental pressure can cause cells to exhibit certain behaviors and can even lead to cell death, and cell damage may affect the results of an experiment. Technical parameters for chip design are selected on the basis of their ability to prevent cell damage, research requirements, and other factors that may improve cell functionality. For example, methods requiring a large number of cell samples would not be suitable for rare cell samples. Samples containing circulating tumor cells are rare. Therefore, the appropriate method would require high-efficiency capture of samples with rare cell subsets (e.g., valve and trap methods). Additional parameters include culture space, observation, downstream analysis, operating time, and throughput. Even when lab-on-a-chip systems with high cell viability are used, cells are affected by the related long-term manipulation. Therefore, a system's operating time must be as short as possible, and the throughput must be high. If these conditions cannot be satisfied, the prevention of cell damage must be prioritized. The difficulty of operating devices and their range of applications must also be considered. Highly complex microfluidic devices are difficult to operate, require high precision during manufacture, and are costly to mass produce. The challenge for a single-cell coculture system is that it must enable the capture, cultivation, and analysis of several types of cells from a single space. The capture process often involves the injection of cells into a chip, the removal of uncaptured cells from the chip, and the injection of another cell type, which is time consuming. An alternate process involves the simultaneous injection of various cells into different channels, the removal of uncaptured cells from the chip, and the connection of the different channels containing the individual cells. This often requires a combination of two or more technologies. For example, traps are used to capture a single cell, and valves are used to seal multiple traps in a single space to form a cocultivation environment for the captured cells. Traps are limited by their structure and flow resistance (this is discussed in detail in Sec. III). A flow channel that is too large reduces the capture efficiency of the trap, and a flow channel structure that is too small does not provide adequate cultivation space. In addition to this problem, researchers should investigate the influence of the contact area in cell–cell interaction,⁹ control the distance and contact area between cells, and incorporate the aforementioned factors in the design of technology to prevent cell damage. Section III introduces existing single-cell coculture chips, describes the use of these technologies in single-cell coculture, and presents an analysis of their advantages and disadvantages to serve as a reference for chip design.

TABLE I.

Characteristics of single-cell capture technologies.

	Microwella	Valvea	Trapa	Dropleta
a. Cell viability	>95%	>90%	75%–85%	>90%
b. Capture efficiency	37%b	>80%	>80%	37%b
c. Throughput (cells/min)	<1 × 106	<1 × 104	<1 × 104	<1 × 106
d. Culture	3–7 days	1–7 days	1–3 days	1–3 days
e. Isolate single-cell directly	No	Yes	No	Yes
f. Downstream analysis	On chip	Collectable	On chip	Collectable
g. Number of cells (106/ml)	0.08–10	0.005–2	0.1–10	1–5
h. Operation time	5–40 min	1–20 min	3–30 min	1–10 min
Reference	19–23 and 25–26	27–31	32–34 and 36–40	41–44

^a All functions and values are from the reference.

^b In cases with no control system or special design elements, the Poisson distribution can be used.

III. ADVANTAGES AND DISADVANTAGES OF SINGLE-CELL CAPTURE TECHNOLOGIES IN MULTIPLE SINGLE-CELL CAPTURE

A. Microwell

The microwell chip uses gravity and cell size to capture cells. The advantages of this technology are that it is simple to manufacture, easy to operate, and has a high throughput for microwell arrays, making it the most widely used single-cell capture technology.¹⁸ Figures 3(a) and 3(b) present examples of the use of microwell arrays to capture cells and analyze thousands of single-cell reactions.^19,20 The high-throughput method produces dense microwell arrays with smaller wells to capture more individual cells, thus increasing capture efficiency. This method does not require any chemical surface modification to trap individual cells in the microwells. However, the cell cultivation area is small, and thus, the culture time and amount of medium usable are limited. Therefore, the cell behavior observation time is insufficient. In a 20 × 20-μm² microwell [Fig. 3(c)], the single-cell capture efficiency decreases as the cell density increases.²¹ Even if cell-specific antibodies are used to increase the surface adhesion for specific cells, the capture efficiency depends on how closely the microwell size matched the cell size;²⁰ and a larger microwell can capture two or more cells at a time. Therefore, to increase capture efficiency and ensure that only one cell is captured per microwell, either the cell concentration or microwell size can be reduced. However, the microwell size can be adjusted for cell lines with a wide range of cell sizes. In such a case, prescreening of cells of a specific size is required, which hinders the effective evaluation of the data distribution of the entire population. Microwells that are too small do not provide sufficient culture space, making long-term observation difficult. Furthermore, high-density cell suspension can cause single cells to stick together and form clusters after dispersion, which reduces the efficiency of single-cell capture. In addition to surface modification methods, different microwell shapes can be used to increase capture efficiency [Fig. 3(d)]. Fluid passing through microwells of different shapes generates various recirculation flow effects, which allows for single cells to enter the microwell recirculation zone, thus increasing single-cell capture efficiency. A microwell in the shape of an equilateral triangle yields the highest single-cell capture efficiency.²² Because size and shape affect cell capture efficiency and culture space, they should be considered in the design of microwells for single-cell capture. The density of single-cell suspension is also crucial; and excessively high cell density reduces single-cell capture efficiency. In addition, a single-cell coculture device must enable the effective pairing of single cells for cocultivation. Individual cells of two types can be separately captured in the same microwell for cocultivation using a coverslip with cytokine-specific antibodies, and the secretory factors produced by the interaction between the cells can be analyzed [Fig. 3(e)].²³ The microwell size must be suitable to hold two cells. To increase capture efficiency, the cell density must be adjusted, which is affected by the probabilistic effect of the Poisson distribution. The expected capture efficiency for two cells is 37%, and that for three cells is 14%.²⁴ Although the capture efficiency is low, the method can be used to effectively isolate each coculture microwell and prevent interference between microwells. Figure 3(f) displays an example of another method of cell cocultivation. With this method, three types of cells are captured in different flow channels of microwells. Then, the flow channel is removed, and a chip is placed in a culture dish for cocultivation.²⁵ The size of the microwell can be modified to suit different cells, and the surface of the channels can be modified to increase capture efficiency for specific cells. This method can be used to cocultivate multiple cells. However, the open microwell space cannot be used to effectively isolate cell secretion factors, which results in the mixing of cell secretion factors from the three cell populations. Under such conditions, tracking the interaction between single cells becomes difficult, and only the averaged results for the three cell types can be observed. To solve this problem, additional flow channels or other components such as valves can be used to narrow specific areas. In the method of cocultivation of multiple cells presented in Fig. 3(g), different microwell sizes are used to hold cells of various sizes for cocultivation.²⁶ In this method, microwells of specific sizes are arranged next to each other. After cells of various sizes are captured in the microwells, they are cocultivated. A limitation of this method is that it cannot be used to effectively pair cells with a difference in size of 2–3 μm, but it can effectively generate pairing combinations at different ratios (e.g., 1:2, 1:3, and 1:4) in a deterministic manner. This method can be improved by decreasing the size of the microwell and isolating and increasing the culture space using a concave coverslip with cytokine-specific antibodies. Section III B provides examples to illustrate the method involving microwells and valves.

FIG. 3.

Multiple single-cell capture using microwell chips. (a) and (b) Single-cell capture using microwell arrays. (c) Surface modification to capture specific single cells. (d) Effect of different microwell shapes on single-cell capture. (e) Cocultivation of single cells in microwells through direct spread. (f) Capture of multiple single cells through different channels and cocultivation. (g) Capture of multiple single cells using different microwell sizes and cocultivation. Image (a) has been reproduced with permission from Tokimitsu et al., Cytom. A 71(12), 1003–1010 (2007). Copyright 2007 International Society for Analytical Cytology. Image (b) has been reproduced with permission from Rettig et al., Anal. Chem. 77(17), 5628–5634 (2005). Copyright 2005 American Chemical Society. Image (c) has been reproduced with permission from Revzin et al., Lab Chip 5(1), 30–37 (2005). Copyright 2005 Royal Society of Chemistry. Image (d) has been reproduced with permission from Park et al., Microfluid. Nanofluidics. 8(2), 263–268 (2010). Copyright 2009 Springer-Verlag. Image (e) has been reproduced with permission from Varadarajan et al., J. Clin. Invest. 121(11), 4322–4331 (2011). Copyright 2011 The American Society for Clinical Investigation. Image (f) has been reproduced with permission from Lin et al., RSC Adv. 6, 75215–75222 (2016). Copyright 2016 Royal Society of Chemistry. Image (g) has been reproduced with permission from Zhou et al., Cell Rep. 31(4), 107574 (2020). Copyright 2020 Elsevier.

B. Valves

Valve chips use pressure to control switches for single-cell capture. Most of these types of microfluidic chips have a multilayer structure of at least two layers: a fluid layer and valve control layer. Because the opening and closing of the valve can be controlled with pressure, fluid movement can be effectively directed to a desired position, and the fluid can be easily isolated. By using a valve, each array unit on a chip can be closed to form an independent space. However, because the valve is controlled through a complex structure, manufacturing multilayered wafers is complicated and time consuming. Valve control also requires the use of a pressure controller. Figure 4(a) presents the single-cell capture process using valves. Opening the valve enables the capture of cells, and closing the valve enables the release of the captured cells.²⁷ Because of the reduction in space, the captured cells are subject to fluid pressure; therefore, the stability of the valve switch and the pressure level must be effectively controlled to prevent cell damage. Figure 4(b) presents a system combining a valve and a trap for cell capture.²⁸ This method uses crossed-control layers of independent trap switches to release individual cells. The advantage of this method is that cells can be selected from the valve chip and that it can be combined with conventional experimental systems for downstream analyses such as Western blot and polymerase chain reaction. The valve chip in Fig. 4(c) combines a microwell and magnetic beads. The cells labeled with magnetic beads are first captured in the microwell, and the microbeads with fluorescent labels are captured in the same microwell. Then, the valve is closed to ensure protein immunity.²⁹ Although this method can be used to repeatedly capture different cells or beads for multiple single-cell cocultivation and has several advantages for downstream analysis, its capture efficiency is affected by the Poisson distribution. In addition, to increase the efficiency of pairing, this method must be combined with other methods. The valve chip in Fig. 4(d) uses a trapping structure with multiple valves to capture single cells. In this method, the valve switch is used to control the direction of fluid flow to capture individual cells in succession.³⁰ The beads in the chip can be replaced with cells, and the same operating principle can be used to increase the number of traps for multiple single-cell cocultivation. Furthermore, a single trap switch can be used to release individual cells for culture observation and downstream analysis in the chamber. However, controlling such as complex multivalve system requires software and pressure control valves. Figure 4(e) presents a method similar to that in Fig. 4(a). In this method, the upper and lower flow channels are controlled by a middle valve to capture single cells from the right inlet. After the remaining cells are removed through the left outlet, the valve is opened to allow the single cells to move to the lower flow channel for analysis.³¹ By increasing the area of the lower flow channel and repeating these steps, cells can be captured individually in the lower flow channel for multiple single-cell cocultivation.

FIG. 4.

Multiple single-cell capture using valve chips. (a) Selective retrieval using a single-cell valve chip. (b) Position is selected by the cross valve. (c) Magnetic beads are combined with valves to capture and isolate individual cells. (d) A hydrodynamic trap combined with valves to capture and isolate individual cells. (e) Use of lateral flow channels and valves to capture individual cells on microwells. Image (a) has been reproduced with permission from Kim et al., Microfluid Nanofluid. 16, 623–633 (2014). Copyright 2013 Springer Nature. Image (b) has been reproduced with permission from Kim et al., Lab Chip 15(11), 2467–2475 (2015). Copyright 2015 Royal Society of Chemistry. Image (c) has been reproduced with permission from Armbrecht et al., Microsyst Nanoeng. 5, 55 (2019). Copyright 2019 Springer Nature. Image (d) has been reproduced with permission from Cheng et al., Nat. Commun. 10(1), 2163 (2019). Copyright 2019 Springer Nature. Image (e) has been reproduced with permission from Kim et al., Microfluid Nanofluid. 13, 835–844 (2012). Copyright 2012 Springer Nature.

In summary, the advantages of valve chips are that they allow for the operation of each unit individually with sufficient culture space, enable the isolation of single cells for downstream analysis, and are highly compatible with other capture methods. The disadvantage is that they require a complex structural design and control systems. Manufacturing valve chips is also more expensive and difficult than manufacturing microwells and trap valves.

C. Traps

Trap chips use physical constraints to limit cells to a special structure for single-cell capture. The most common trap chip is the U-trap. Trap chips are usually combined with hydrodynamics to increase capture efficiency. Hydrodynamic capture is achieved using two streams of fluid with different resistances to guide single cells into the trap structure. Because the flow resistance of the trap structure is lower than that of the bypass channel, fluids tend to move toward the trap structure, resulting in the entrapment of single cells in the structure. Once the trap structure is blocked by a single cell, its fluid resistance increases substantially, which causes the fluid to change direction and move to the bypass channel. The remaining single cells tend to move toward the bypass channel and are captured in the trap structure. The capture efficiency of this method can be increased by calculating resistance and performing a simulation analysis to optimize the trap and flow channel structure and overcome Poisson distribution-related limitations. The capture performance of this deterministic method can be controlled through precise calculation, and the number of cells required for a sample can be reduced substantially. Figure 5(a) illustrates the structure of this single-cell capture method.³² This method uses a hook-shaped trap to capture single cells and high-density cell suspension to prevent issues from occurring during capture. Unlike in the deterministic method, the hook structure used to capture cells in this method is close to the chip microchannel wall. Consequently, the fluid resistance in the narrow flow path between the hook and the wall is higher than that in the wide flow path on the side without the hook. This makes the fluid resistance of the flow channel on the wide side lower than that of the flow channel on the narrow side, which causes cells to approach the channel on the wide side. Therefore, capture efficiency can be increased by increasing cell density. In this method, hooks are used to capture different types of single cells in a fixed position, and the hook channel is then removed to coculture the captured cells. The distance between the hooks can be modified depending on the requirements of an experiment to control the distance between individual cells and allow them to make contact. Figures 5(b) and 5(c) present the hydrodynamic method of capturing different types of cells for cocultivation^33,34 and cell fusion.³⁵ This method uses a constricted channel structure and dense array to increase cell capture efficiency. Therefore, the density of the single-cell suspension is crucial in preventing cell clusters from blocking the flow channel. Figure 5(d) presents the same method with an increased number of traps to capture multiple single cells of the same type for coculture and interaction analysis.³⁶

FIG. 5.

Multiple single-cell capture using trap chips. (a) Block-cell printing using a hook-shaped trap. (b) Single-cell coculture using a cellular valving trap. (c) High-throughput deterministic single-cell pairing array. (d) Multiple single-cell immobilization chips. (e) Two-phase flow isolation for single-cell pairing and coculture. (f) Hydrodynamic single-cell pairing through attachment of cells to microwells. (g) Loading dock system for multiple single-cell capture. (h) Hydrodynamic shuttling method for deterministic multiple single-cell capture. Image (a) has been reproduced with permission from Zhang et al., Proc. Natl. Acad. Sci. U.S.A. 111(8), 2948–2953 (2014). Copyright 2014 National Academy of Sciences. Image (b) has been reproduced with permission from Frimat et al., Lab Chip 11(2), 231–237 (2011). Copyright 2011 Royal Society of Chemistry. Image (c) has been reproduced with permission from Dura et al., Nat. Commun. 6, 5940 (2015). Copyright 2015 Springer Nature. Image (d) has been reproduced with permission from Tang et al., Anal. Chem. 92(17), 11607–11616 (2020). Copyright 2020 American Chemical Society. Image (e) has been reproduced with permission from Chen et al., Lab Chip 14(16), 2941–2947 (2014). Copyright 2014 Royal Society of Chemistry. Image (f) has been reproduced with permission from Hong et al., Integr. Biol. 4(4), 374–380 (2012). Copyright 2012 Oxford University Press. Image (g) has been reproduced with permission from Li et al., Adv. Biosyst. 1(10), e1700085 (2017). Copyright 2017 John Wiley and Sons. Image (h) has been reproduced with permission from He et al., Lab Chip 19(8), 1370–1377 (2019). Copyright 2019 Royal Society of Chemistry.

These methods lack sufficient cultivation space, require additional channels, and must be combined with other methods for long-term cultivation and observation. Figure 5(e) presents the use of an additional semipermeable membrane and media exchange layer for long-term cell culture observation.³⁷ This chip consists of traps in the same flow channel, and each trap can capture two cells of the same or different type. Because the successful pairing of cells of different types is random, the efficiency of cell pairing in this method is limited. Figure 5(f) presents the use of the deterministic hydrodynamic method to increase the capture area, thus allowing for efficient culture observation.³⁸ This chip uses hydrodynamic properties to capture individual cells at the position of a trap, after which the cells can attach to the culture area in front of the trap. After the cells leave the trap, the chip captures individual cells moving toward the trap in succession. This operation is repeated to capture multiple single cells for cocultivation. The limitation of this chip is that the second capture can be conducted only when all cells from the first capture are attached to the culture area, which makes the simultaneous coculture of individual cells impossible. Figure 5(g) presents a trap structure that uses various pressure levels to organize different types of cells.³⁹ This structure comprises an outer trap and an inner arrangement trap. The inner arrangement trap has openings on the side to ensure that fluid can flow out and that a second cell can enter the trap, thus preventing the first cell from being subject to excessive pressure. Cells are first captured in the outer trap by applying low pressure. Then, the pressure is increased to push the cells into the arrangement trap. This operation is repeated to capture multiple single cells. This chip designed a 3D chamber on top of the arrangement trap for long-term culture observation. Figure 5(h) presents the use of a hydrodynamic method to capture individual cells in the trap and push them into the large culture chamber with reverse flow.⁴⁰ The large culture chamber can cause the flow rate of the cells to decrease rapidly after they enter the chamber, thereby allowing for individual cells to remain in the culture chamber. The chip can repeatedly capture individual cells in the large culture chamber and allow for cultivation and the observation of the interaction between multiple single cells. Because of the resistance, the operating time must be shortened to prevent damage to cells in the trap structure. Hydrodynamic principles cause the limited channel space to be easily blocked by unevenly dispersed cell clusters, which affects capture efficiency.

In summary, trap chips combined with hydrodynamics can effectively resolve the Poisson distribution problem, and the additional 3D chamber can enable long-term culture observation. However, cell damage from fluid pressure during operation should be prevented. Blockage of the channel by cell clusters should also be considered.

D. Droplets

Droplet chips use oil to shear water and form a stable water-in-oil droplet. By controlling the cell density in the water phase, single cells are encapsulated in droplets with the oil phase liquid. The probability of capture efficiency follows the Poisson distribution. This method can control the flow rate of the two-phase liquid and produce a large number of droplets in a short amount of time. Each single-cell droplet is an independent unit with no mutual interference. Figure 6(a) presents a typical droplet chip. Single cells are encapsulated in droplets through the shearing of the water phase containing single cells by the oil phase.⁴¹ The cell density of the water phase strongly affects single-cell capture efficiency. The oil phase material should be transparent and must be able to facilitate the delivery of gas into droplets to enable optical observation, analysis, and cell culture. More samples can be encapsulated in droplets by increasing the number of inlets and designing an array structure to group droplets together [Fig. 6(b)].⁴² This method can be used to encapsulate drugs or coculture cells. The array structure is conducive to observation and analysis. Droplets without this array structure usually mix together, making tracking and observation difficult. Figure 6(c) presents a structure with large microwells for capturing single-cell droplets with large diameters and small microwells for capturing droplets with smaller diameters. A demulsifier combines the two droplets for single-cell cocultivation analysis.⁴³ This chip uses optical signal sorting technology, which can eliminate the Poisson distribution problem in single-cell capture. In addition, the chip can be flipped to allow the droplets to be suspended and released for downstream analysis. The droplet chip in Fig. 6(d) combines microwells, valves, and hydrodynamics to generate, store, and merge droplets of different sizes.⁴⁴ This chip can also generate droplets of different volumes through the adjustment of valve opening time and inlet pressure (droplet generation unit). The combination of the hydrodynamic microwell structure and valve can allow for droplets to flow into the microwell when the valve is closed and move via the bypass channel to the next microwell (FIFO storage unit, the array operates in a first-in, first-out manner) when the valve is open. The volume of the microwell structure affects the flow rate in the microwell; and a smaller volume increases the flow rate, which makes it difficult for droplets to remain in the microwell. By contrast, a larger volume considerably reduces the flow rate and ensures that the droplet remains in the microwell. This method has higher error tolerance for valve control accuracy. Droplets are merged by the hydrodynamic microwell structure with a valve (merging unit). The valve is first closed to introduce the droplet into the microwell, opened to receive the next droplet, and finally closed to ensure contact between the first and the second droplet for merging. Thus, various combinations of cells can be generated [Fig. 6(e)]. However, this chip lacks the deterministic function for cell encapsulation; therefore, capture efficiency is limited by the Poisson distribution. Therefore, both the droplet and the valve methods rely on the control system to regulate the pump flow rate. An array structure of fixed droplets is essential to track and analyze individual droplets. In addition, droplets can only be used for suspension cultures, and adherent cells cannot be cultured in droplets. Therefore, the droplet capture technology cannot be applied to cell migration research.

FIG. 6.

Multiple single-cell capture using droplet chips. (a) Water-in-oil emulsion droplets for single-cell capture. (b) Droplet array for single-cell drug screening. (c) Deterministic droplet sorting and pairing chip. (d) Comprehensive multiple single-cell capture chip. (e) Results of controlling cell numbers. Image (a) has been reproduced with permission from Clausell-Tormos et al., Chem. Biol. 15(5), 427–437 (2008). Copyright 2008 Elsevier. Image (b) has been reproduced with permission from Sarkar et al., Lab Chip 15(23), 4441–4450 (2015). Copyright 2015 Royal Society of Chemistry. Image (c) has been reproduced with permission from Chung et al., Lab Chip. 17(21), 3664–3671 (2017). Copyright 2017 Royal Society of Chemistry. Images (d) and (e) have been reproduced with permission from Babahosseini et al., Lab Chip. 19(3), 493–502 (2019). Copyright 2019 Royal Society of Chemistry.

IV. SUMMARY AND OUTLOOK

Conventional multiple single-cell coculture technologies cannot be used to effectively identify and analyze differences among individual cells in a heterogeneous sample. Heterogeneous cells often play a key role in the function of the entire cell population. To effectively identify and analyze differences in individual cells, composite microfluidic devices with high capture efficiency, culture environment control, cell tracking, and detection and analysis capabilities are essential. Because of technological limitations, few studies have investigated multiple single-cell coculture chips. This paper provides a brief introduction to microfluidic technologies for single-cell capture and compares their advantages and disadvantages. Figure 6(d) presents the optimal chip design for multiple single-cell cocultivation through microfluidic technology. The advantages of droplet capture are its high throughput, low cell damage, and efficient isolation and collection. A hydrodynamic trap structure can be added to a droplet capture system to achieve deterministic capture efficiency and remove cells from droplets for culturing in separate microwells. Downstream detection and analysis are essential to understand the interactions between cells, such as protein–protein interaction and single-cell RNA sequencing. Higher accuracy can be achieved by comparing the mRNA and protein expression levels of single cells.⁴⁵ Some microfluidic chips for single-cell collection not covered in this paper can also be used.^46,47 Downstream analysis can also be directly performed on the chip,^48–50 but the issues of multifield technology integration and the complicated chip manufacturing process remain. Nevertheless, the rapid development of microfluidic technology may allow for the underlying mechanisms of cell–cell interaction to be determined. The successful design of such a cell capture technology would represent a substantial breakthrough for the fields of cancer treatment, biologics, tissue engineering, and regenerative medicine. The development and manufacture of microfluidic chips require the integration of microprocessing technology with aspects of chemical engineering, materials engineering, biomedical engineering, microelectromechanical engineering, programming, and life sciences. The successful integration of these technologies can help realize the idea of lab-on-a-chip.

AUTHOR DECLARATIONS

Conflict of Interest

The authors have no conflicts to disclose.

Author Contributions

C.-K.H. wrote the manuscript with the support and guidance of C.-H.H. Both C.-K.H. and C.-H.H. contributed to the final version of the manuscript.

DATA AVAILABILITY

Data sharing is not applicable to this article as no new data were created or analyzed in this study.