Recent advances in single cell manipulation and biochemical analysis on microfluidics

Dan Gao *ab, Feng Jin b, Min Zhou a and Yuyang Jiang *ac
aState Key Laboratory of Chemical Oncogenomics, Key Laboratory of Chemical Biology, Graduate School at Shenzhen, Tsinghua University, Shenzhen 518055, P.R. China. E-mail: gao.dan@sz.tsinghua.edu.cn; jiangyy@sz.tsinghua.edu.cn
bDepartment of Chemistry, Stanford University, Stanford, California 94305, USA
cSchool of Pharmaceutical Sciences, Tsinghua University, Beijing 100084, P.R. China

Received 27th June 2018 , Accepted 14th September 2018

First published on 14th September 2018


Single cell analysis has become of great interest with unprecedented capabilities for the systematic investigation of cell-to-cell variation in large populations. Rapid and multi-parametric analysis of intercellular biomolecules at the single-cell level is imperative for the improvement of early disease diagnosis and personalized medicine. However, the small size of cells and the low concentration levels of target biomolecules are critical challenges for single cell analysis. In recent years, microfluidic platforms capable of handling small-volume fluid have been demonstrated to be powerful tools for single cell analysis. In addition, microfluidic techniques allow for precise control of the localized microenvironment, which yield more accurate outcomes. Many different microfluidic techniques have been greatly improved for highly efficient single-cell manipulation and highly sensitive detection over the past few decades. To date, microfluidics-based single cell analysis has become the hot research topic in this field. In this review, we particularly highlight the advances in this field during the past three years in the following three aspects: (1) microfluidic single cell manipulation based on microwells, micropatterns, droplets, traps and flow cytometric methods; (2) detection methods based on fluorescence, mass spectrometry, electrochemical, and polymerase chain reaction-based analysis; (3) applications in the fields of small molecule detection, protein analysis, multidrug resistance analysis, and single cell sequencing with droplet microfluidics. We also discuss future research opportunities by focusing on key performances of throughput, multiparametric target detection and data processing.


1. Introduction

Cells are the basic structural and functional units of living organisms. Cells derived from a mother cell or from the same type of cells under the same physiological conditions or external stimuli exhibit cell-to-cell differences including the size, growth rate, morphology, and different expressions of biological molecules.1–3 In terms of various tumor-related research studies, cellular heterogeneity has a close relationship with tumor metastasis, drug resistance, and clinical diagnosis. For example, different responses of individual cells to drugs cause the emergence of drug-resistant cells, but only a small percentage (0.3%) of these cells have the ability for tumor recurrence.4,5 Nowadays, the heterogeneity of molecules in individual cells has become a huge obstacle to the development of effective cancer target treatments.6 However, the average data obtained from conventional bulk cell assays would neglect the differences in intracellular molecular expression among individual cells, causing the loss of important biological information. Therefore, analysis of cellular components at the single cell level is not only an effective way to probe cell heterogeneity, but also helps us comprehensively understand the behaviors of tissues, organs, and even the whole living organism.

Compared with bulk cell analysis, single cell analysis poses some significant challenges with respect to the sizes of cells and the concentrations of cellular components. Owing to the small size of the majority of cells, manipulation becomes difficult when using traditional sample preparation strategies.7 In addition, most of the intracellular components are present in small concentration and span a wide range of concentration regimes, which further demand highly sensitive and specific detection methods. Earlier methods for single cell analysis mainly relied on flow cytometry or laser scanning cytometry by rapidly screening fluorescently labeled cells in a flow.8,9 Take flow cytometers as an example; although they are automatic and capable of multiple detections, they are bulky, mechanically complicated and demanding in terms of professional training. In addition, they can hardly detect particles smaller than 0.5 μm in diameter based on light scattering. Recent developments in microfluidic techniques provided promising tools for single-cell analysis.10,11 Microfluidics provides several benefits over conventional techniques. Most importantly, some of the benefits can be integrated into one platform. Firstly, the structures and functions of the microfluidic chip can be flexibly designed to fulfill the demands of single cell analysis. Secondly, miniaturized microfluidic systems work with picoliter to nanoliter volumes of solution, which help to reduce sample loss and decrease dilution, resulting in highly sensitive assays. Thirdly, their parallelization ability makes it possible for parallelized analysis. Fourthly, multiple detection techniques can be easily combined with microfluidic chips, for example, optics, electricity, mass spectrometry, etc. for online and real-time assays. Last but not least, automation through compact integration of multifunction units into the same chip can greatly prevent measurement errors generated from human operations. Due to its advantages, microfluidics has become a popular technology for various applications in life sciences,12–14 and it has been demonstrated to be a powerful tool for highly efficient single cell analysis. Until now, different strategies based on microfluidic technology have been developed for single cell capture, such as microwells,15–17 droplet-based methods,18,19 microvalves,20 optical tweezers,21 microtraps,22 and flow cytometric methods.23 Various detection methods can be combined with microfluidic systems for single cell analysis; the most widely used online detection techniques are fluorometry24 and spectroscopy.25 Fluorescence microscopy technique is capable of analysing the preselected molecules marked by molecular probes or reporters, but the types of detected components are limited. By comparison, mass spectrometry (MS)-based single cell analysis methods have attracted increasing attention due to the advantages of label-free detection and their ability to detect unknown molecules. So far, single cell analysis based on multi-technical combinational microfluidics has revolutionized both fundamental and applied research fields, such as drug and toxicity screening,8,26 clinical studies,27 stem cell research,28 gene expression analysis,29 (whole) transcriptome profiling,30,31 proteomics analysis32 or enzyme activity,33 and metabolite analysis.34,35

Recently, many reviews on single cell manipulation, single cell sample pretreatment, different kinds of detection methods, and various practical applications in the microfluidics research field have been reported.36–39 In this review, we mainly focus on some of the strategies based on microfluidic technology for single cell manipulation and biochemical component analysis of individual cells for biology and medicine applications over the past three years. We highlight technical improvements or emerging techniques that are promising for future developments, which are discussed in terms of single cell manipulation including microwells, micropatterns, droplets, traps, and flow cytometric methods. Analytical techniques focus on the most popularly used detection methods based on fluorescence, mass spectrometry, electrochemical, and polymerase chain reaction (PCR)-based analysis. Finally, we also discuss the future directions in the improvement of microfluidic techniques and data processing methods for more comprehensive single cell analysis.

2. Single cell manipulation

The first step in single cell analysis on microfluidic devices is to isolate individual cells and to dock them in a desired location for further treatment. Conventional methods contain several steps including serial manual dilution by micropipetting, fluorescence activated cell sorting using a flow cytometer40 or laser capture microdissection through direct microscopic visualization.41 Some disadvantages are that they mechanically complicated and cumbersome. Microfluidics is regarded as a powerful tool for single cell analysis as it helps build desired small devices for cell manipulation and can be combined with many analytical methods.37,42 Various recent microfluidic developments related to highly efficient single cell manipulation techniques will be reviewed and discussed.

2.1 Microwells

Parallel immobilization of single cells with high occupancy rates is a prerequisite to investigate the response of single cells to environmental factors or drug treatment. One commonly used method to achieve this goal is to design cell-sized microwells or chambers to dock cells by gravity-induced sedimentation. When cells have a higher density than the surrounding medium, the cells can be passively captured with a higher capture efficiency. In addition, the microwell's geometry, size, depth, and material properties can be easily adjusted to capture the organism of interest.43 However, trapping of single cells by sedimentation needs a relatively longer time scale and the capture efficiency is usually not very high, ranging from 2.6% to 39%.44 Therefore, it is necessary to develop new methods for fast single cell capture with high efficiency.

Making some modifications in microwell structures is a direct and effective way to improve single cell capture efficiency. However, it may take a longer time to reach high capture efficiency. In order to overcome some drawbacks, except for structure modifications, other external operations like those aided by vacuum or centrifugation can be simultaneously applied to single cell immobilization and capture. Some successful examples have been reported in recent years. For example, Terstappen et al. designed a self-seeding microwell chip containing 6400 microwells with a single 5 μm pore in the bottom of each microwell for highly efficient single cell capture (Fig. 1A).45 In order to realize fast single cell capture, the microwell chip was degassed in a vacuum chamber and a small negative pressure of 10 mbar was applied across the microwells. Under the hydrodynamic forces, a single cell could be easily dragged into the pore when the fluid entered the microwell. With a slight structural modification of the microwells, the single cell capture efficiency could reach up to 67%. Wu et al.46 used a truncated cone-shaped microwell array chip to trap single cells with the assistance of centrifugation, and a single-cell occupancy of approximately 90% could be achieved within a few seconds (Fig. 1B). A key improvement of this paper is that the designed truncated cone-shaped microwell with the opening diameter smaller than its bottom can effectively prevent significant loss of cells during exposure to various chemicals. Lee et al.47 also used centrifugation to deposit a single cell into the designed microwell array. A higher capture efficiency of 99.93% could be obtained by loading an appropriate cell concentration which is guided by theoretical calculation. In order to improve single-cell capture efficiency without compromising on loading time, dielectrophoresis (DEP) is also exploited to generate a focusing force that helps in single cell patterning in the microwells.48 Although microwell-based microfluidics offers several advantages such as combining with additional operations to exchange solutions without pipetting, their operations still involve several manual handling steps. The majority of the established microwell-based platform are very simple, with only one fluid channel on top of the microwells to allow paracrine communication between cells generated during experiment, which will interfere with accurate multi-parameter detection in single cells. To solve this problem, Garcia-Cordero et al.49 developed a multilayered microfluidic device with a microvalve sitting on top of the microwells to create an independent microenvironment and to avoid paracrine communication between wells. Another key advantage was the very low shear stress (<0.02 Pa) inside the wells which could keep the original biological behaviors of cells. Due to these significant features, the device could be applied in cell biology and drug discovery that require long-term observation of morphological and structural differences in single cells.


image file: c8an01186a-f1.tif
Fig. 1 (A) Self-seeding microwells with a pore in the center. Reprinted with permission from J. F. Swennenhuis, et al., Lab Chip, 2015, 15, 3039–3046. Copyright 2015 the Royal Society of Chemistry. (B) Schematic illustration of a polydimethylsiloxane (PDMS) chip with a truncated cone-shaped microwell array and centrifugation-assisted single-cell trapping. Reprinted with permission from L. Huang, et al., Anal. Chem., 2015, 87, 12169–12176. Copyright 2015 American Chemical Society.

So far, several microfabrication technologies have been developed to fabricate microwell arrays for high throughput experiments, for example, soft lithography,50 biochemical patterning,51 direct printing52 and laser processing.53 Soft lithography is regarded as the most popular and widely used approach, but it is not the best choice because of complex fabrication strategies and the requirement of special equipment and cleanrooms.54 In terms of patterning, printing or laser processing methods for fabricating microwell arrays, the fabrication steps are simpler than soft lithography but additional chemical modification or replica molding is indispensable. In order to overcome the disadvantages of time-consuming, cumbersome and current expensive approaches, simpler and more efficient methods are still needed. Liu et al. developed an easy-to-use method to fabricate a poly(ethylene glycol) diacrylate (PEGDA) hydrogel microwell chip by using a digital micro-mirror device printing system in which three different dimensions including a single cell (1D), a cell monolayer (2D) and a cell spheroid (3D) can be easily formed in one device.55 A digital mask was employed in the microwell manufacturing process based on shadowed light instead of a conventional physical mask. The shape, dimensions, and distribution of microwell arrays can be well controlled via the mask-free hydrogel fabrication system. In addition, the constraining hydrogel film could be peeled off from the glass substrate for further cell analysis. Due to recent developments in fabrication techniques, materials science and other operation technologies, the microwell approach has been widely used for high-throughput assessment of gene expression, protein and enzyme assays, cell secretion, and drug screening.

2.2 Micropatterns

Micropatterning is another frequently used method to form single cell arrays using chemical surface patterns with cell adhesive spots surrounded by cell-repellent surfaces. The micropatterning of cells with different shapes has been widely used for the study of cell behaviors and cell–material interactions, such as cell shape, polarity, migration and division.56,57 The commonly used materials for the formation of adhesive regions are biomimetic materials, such as fibronectin, laminin, vitronectin, and collagen, whereas PEG and other containing hydrophilic polymers are often used for cell-repellent surface modification. There are several commonly used methods to produce chemical surface patterns, for example, microcontact printing,51 inkjet printing,52 or photopatterning.58 However, attaining a homogeneous distribution of cells over the entire patterned area is difficult in microfluidic systems because of non-specific absorption on the cell-repellent area.

Over the past three years, there was no significant progress in the application of micropatterning methods to single cell arrays formation. However, referring to cancer cell analysis, especially for circulating tumor cell (CTC) isolation, a high capture efficiency is extremely important because of its very low concentration. Recently, micropatterning of chemical linkers like aptamers or antibodies to coat the surface of the microchannel or micropillars is popularly used to trap free-floating single cells from a biological sample. The capture efficiency could be largely increased due to the increasing contact between cells and capture sites. Yang's research group59 coated the hydrodynamically optimized triangular pillars with an anti-EpCAM antibody to capture CTCs from the blood samples. The authors rotated the triangular micropillars to generate slopes and a gradient of hydrodynamic forces following the slope, which will provide lower shear stress to increase the contact time between CTCs cells and immunodecorated micropillars. Under the optimized conditions, CTCs with different antigen expression levels could be efficiently captured with 92% capture efficiency and 82% purity. Nonetheless, most of the reported micropatterning methods suffer from the drawback that cells cannot be released from the initial capture site for downstream biological analysis. On the basis of this problem, Revzin et al.60 used a photodegradable hydrogel (photogel) functionalized with antibodies to capture and release target cells. Leukocyte-specific antibodies were printed onto the photogel substrates to isolate human CD4 or CD8 T-cells from a peripheral blood mononuclear cell suspension, and then the desired cells were released by UV-induced photo-degradation. The retrieved CD4 and CD8 T-cells revealed approximately 95% purification by flow cytometry analysis.

2.3 Droplets

Similar to the encapsulation of individual cells in different shapes of microwells, cells can be enclosed in femtoliter to nanoliter droplets to form separate microchambers for individual reaction with each cell placed in its own separate droplet. The risk of cross-contamination can be largely decreased and the solution inside the droplet can facilitate rapid mixing, thereby minimizing sample dilution. Due to the above advantages, droplet-based microfluidics is of particular interest and has been widely used in cell biology, clinical research and drug discovery.

Typically, highly monodisperse droplets can be generated using immiscible aqueous and oil phases on T-junctions and flow focusing geometries on microfluidic channels. Two-phase emulsions such as water-in-oil (W/O) or oil-in-water (O/W), as well as more complex multiple-phase emulsions have been generated from these channel geometries.61 The size of the droplets and the rate of droplet formation can be adjusted by a set of parameters including channel dimensions, interfacial tension, flow rates, and viscosities.62,63 In addition, droplets can be generated in high frequency (Hz to kHz), which allows for massively parallelized studies on single cells. Although the droplet-based microfluidics face a low single cell encapsulation rate, the total number of captured single cells as well as the throughput are still much higher compared to other techniques. For example, a flow focusing capillary microfluidic device was assembled by Liu's group recently for high throughput generation of microdroplets.64 This method enabled the monitoring of MMP9 enzymatic activities at the single cell level in several minutes. By optimizing the cell loading concentrations, the maximum single cell encapsulation efficiency was around 25% with a droplet generation frequency of 2.4 × 103 droplets per min. However, the channel-based microfluidics is more likely to clog when a higher cell concentration is introduced. Digital microfluidics (DMF), a channel-free microfluidic technology, handles small droplets of liquid on planar surfaces, which can help overcome the clog problem. Lammertyn et al. presented an electrowetting-on-dielectric-based digital microfluidic (DMF) platform to monitor single yeast cell responses on antifungal drug treatment by time lapse fluorescence microscopy in a high throughput.65 The biggest advantage of the DMF platform was that only surface tension and physical forces generated by moving the droplet were used to capture single cells. Therefore, cell damage could be largely avoided.

Except for high throughput formation of droplets by continuous flow-based droplet systems, the droplet can also be generated by two dimensional (2D) static droplet array systems. They usually combine with microchips with microwells66,67 or micropatterning68 arrays to control the droplet volume. However, most of the droplet-based microfluidic systems met the problem of adding reagents into the generated droplets. Fang's group69 developed a solid pin-based droplet system, which used the solid pins to dip and deposit liquids and moved the two-dimensional oil-covered hydrophilic pillars for miniaturized liquid–liquid reactions and assays (Fig. 2A). Using the “dipping-depositing-moving” manipulation procedures, reagents could be easily added to the sample droplet on the hydrophilic pillar. The authors successfully used it to measure the kinetic parameters for matrix metalloproteinase-9 in picoliter droplets and detect the enzymatic activity in single cells. However, this technique suffered from limited throughput of droplet generation and manipulation, and limited sample pretreatment procedures. The research group later developed a nanoliter-scale oil–air–droplet (OAD) chip combined with droplet-based microfluidics, which could realize multistep complex sample pretreatment and injection for single cell proteomic analysis (Fig. 2B).70 Because all sample pretreatment and injection procedures could be performed in the nanoliter-scale droplets, sample loss can be minimized to increase single cell analytical sensitivity. Another key advantage of this work was the development of a self-aligning monolithic device coupled with the OAD chip to assist the direct injection of the droplet sample into a capillary liquid chromatography (LC) column for further LC-MS/MS analysis.


image file: c8an01186a-f2.tif
Fig. 2 (A) Setup of the solid pin-based droplet system. Reprinted with permission from X.-L. Guo, et al., Anal. Chem., 2018, 90, 5810–5817. Copyright 2018 American Chemical Society. (B) Schematic diagrams of the setup of the nanoliter scale oil–air–droplet (OAD) chip and the self-aligning monolithic device and the entire procedure of sample pretreatment and injection for single cell proteomic analysis. Reprinted with permission from Z.-Y. Li, et al., Anal. Chem., 2018, 90, 5430–5438. Copyright 2018 American Chemical Society.

2.4 Traps

Trapping of single cells at fixed positions by active or passive capture strategies has been widely implemented in microfluidic systems. It is a necessary step to maintain the cells for a long period and thus to facilitate cell biology analysis. There are several techniques commonly used to direct cells to specific locations through both contact and noncontact methods, such as hydrodynamic, mechanical, electrical, magnetic, optical or acoustic forces.

By far, hydrodynamic flow is the most commonly used mechanism for trapping cells. In general, microfluidic systems are used to design side channels to the main transport channel, where the cells can be suctioned into the small side channel by focusing flow. The traps can be designed into various shapes like dams, weirs and holes. For example, a modular single-cell pipette (mSCP) consisting of a SCP-Tip, an air-displacement pipette (ADP), and ADP-Tips was developed by Qin's research group to isolate single cells from the cell suspension.71 SCP-Tip is the key part of mSCP, which contains a Y-bifurcation microchannel, a hook for fluid pressure transmission and single cell capture by a hydrodynamic trap. Compared to the current SCP system, the mSCP offers three obvious advantages. Firstly, it allows for more convenient operation and gentle pressure control by a common ADP. Secondly, the SCP-Tip containing a hydrodynamic trap could achieve 100% single cell isolation from the cell suspension. Thirdly, this technology has the ability to operate at relatively low cell concentrations. However, this technique depends on manual operations that cannot fulfil the requirement for high throughput single cell analysis. Aiming at higher throughput, a two-layered microfluidic platform containing a flow channel in the upper layer and a mini-well array on a 76 × 26 mm glass slide with U-shaped hydrodynamic traps for single cell capture in the bottom layer was reported.72 As shown in Fig. 3A, two different flow channels, a serpentine channel and a tree-pattern channel, were designed separately for trapping and culture of adherent cells and non-adherent cells. The trapping efficiency could reach up to 90% through the incorporation of a U-shaped hydrodynamic trap with a notched microchannel on the downstream edge of each well. The medium can be automatically perfused for long-term cell culture and continuous time-lapse imaging. However, retrieval of captured single cells, as an important consideration for further biological analysis, cannot be realized on this microfluidic system. To fulfil this requirement, a serpentine shaped microchannel with a linear array of hydrodynamic capturing sites and filtering structures was fabricated recently to isolate, capture and retrieve the individual cells.73 As shown in Fig. 3B, the incoming cells initially occupy the trapping channels by focusing structures, additional cells diverted the flow through the bypass pathway and positioned in the downstream traps, a constricted conduit of sub-cellular dimensions. The trapping scheme ensures a near-perfect trapping of single-cells. The number of parallel circuits could be easily increased or decreased to adapt to the necessarily processed cells. However, the mechanical trap techniques like hydrodynamic focusing usually generate mechanical stress on cells, which will deform cells and affect the physiological function of cells. Recently, Gracias et al. developed an optical transparent mechanical trap (MT) array that can capture and encapsulate individual cells without any perturbation, and measured the multiplex 3D surface imaging of analytes using surface-enhanced Raman spectroscopy.74 The MT uses the energy released from differential residual stress in patterned nanoscale bilayers to capture cells. As shown in Fig. 3C, four optically transparent arms of the MT could fold by tailoring the thin film stress, so that the cells can be trapped without perturbation. In order to reduce the interference from background Raman signals, SiO and SiO2 were used as transparent arms on a quartz substrate, which further enhanced the spectroscopic signal.


image file: c8an01186a-f3.tif
Fig. 3 (A) Hydrodynamic trapping of single cells in different microfluidic geometries for the trapping and culture of adherent and non-adherent cells. Reprinted with permission from H. Chen, et al., Lab Chip, 2015, 15, 1072–1083. Copyright 2015 the Royal Society of Chemistry. (B) Schematics of the microfluidic circuit and workflow for single cell trapping, encapsulation and retrieval. Reprinted with permission from M. Sauzade, et al., Lab Chip, 2017, 17, 2186–2192. Copyright 2017 the Royal Society of Chemistry. (C) Mechanical trap surface-enhanced Raman spectroscopy for 3D surface molecular profiling of single live cells. Reprinted with permission from Q. Jin, et al., Angew. Chem., Int. Ed., 2017, 56, 3822–3826. Copyright 2017 Wiley Online Library.

Except for these passive techniques, a number of active techniques like optical, magnetic, and acoustic forces are also employed to manipulate and pattern particles and cells. Since the introduction of optical tweezers in 1986, it has become one of the most powerful tools to manipulate micrometer-sized particles and cells with high precision. They have been widely used in the research fields of biology, physics, chemistry and nanoscience. Besides, optical tweezers can combine with some detection methods like Raman spectroscopy to trap individual cells75 or locate surface-enhanced Raman spectroscopy (SERS)-active microspheres onto the surface of cells of interest76 for in situ SERS-based intercellular analysis, which will open a new route to single cell analysis with high spatial and temporal precision. However, some of the optical tweezers are limited to single-beam single-trap systems, leading to low throughput analysis. Recently, Li et al. assembled microlenses on an optical fiber probe to produce a parallel photonic nanojet array, which avoided the use of elaborate nanostructures and bulky optical systems.77 The system enabled the trapping and detection of multiple nanoparticles and subwavelength cells at low optical power, high throughput, single-nanoparticle resolution and high selectivity.

2.5 Flow cytometric methods

Flow cytometry is an effective approach to sort and quantify the cell size and surface markers of single cells by laser induced fluorescence technology.39 In commercial instruments, fluorescently labeled cells in a single aqueous phase are injected into a flow cytometer, hydrodynamically focused on the center fluid channel by a sheath flow and continuously passed through the laser detection area for fluorescence analysis. To confine the center flow stream to be small enough for single cells ananlysis, various microfluidic focusing systems have been developed to ensure the performance of microflow cytometers, such as two-dimensional (2D), three-dimensional (3D), dielectrophoresis and acoustic.

Compared to other microfluidic focusing methods, the advantage of 2D focusing is that it is simple and it is the most commonly used method. For example, Li et al. presented a microfluidic flow cytometer containing a constriction channel with a cross-sectional area smaller than biological cells, so that single cells could be squeezed inside the channel and analyzed.78 Du's research group developed a two-layer microfluidic channel containing a serpentine gas channel and a straight cell channel separated by a PDMS film for microflow.79 Single cells could be manipulated and non-invasively measured at a controlled oxygen level. Zhao et al. presented a novel microfluidic impedance cytometer containing a crossing constriction channel as the flow channel with a pneumatic controller to aspirate single cells.80 Because a much higher pressure can be used to aspirate cells in the microchannel, leading a significant improvements in throughput. The chip design for 3D microfluidic focusing systems is more complicated than the 2D methods because sample flow is focused in both the horizontal and vertical directions. Therefore, the result will be more accuracy and reliable. Acoustic actuation is an effective strategy for sample transport at the microscale. It has significant advantages like a wide operating frequency range and good biocompatibility. Recent research studies found that acoustic actuation can be integrated into microfluidic systems for efficient sample focusing. For example, Huang's research group integrated surface acoustic waves into the microfluidic device, where an array of 3D trapping nodes could be generated for the trapping and manipulation of single cells and particles.81 The 3D acoustic tweezers has the potential to be applied in a fluorescent-activated cell sorter. Neild's research group systematically investigated the relationship between the cell viability and acoustic power.82 They found that most lymphocytes will quickly lyse when the acoustic power is higher than 570 mW while below it, the power will keep cells viable for a long period, and the lysis threshold powers are different for different cells types. In order to pattern one cell per acoustic well rather than clumps, the ratio of acoustic wavelengths to diameters was also investigated.

3. Detection methods for single cell analysis

With the development of novel analytical techniques as well as microfabrication technologies, many microfluidic systems have shown a strong potential to manipulate single cells and have allowed for complex multiparametric assessment, which will help us improve our knowledge and strategies for cell biology analysis. To the best of our knowledge, a single cell contains tiny concentrations of biomolecules with a wide distribution range, which brings a great challenge for single cell component analysis. Therefore, detection methods with high sensitivity are necessary. In the following, we will highlight recent advances in the most commonly used methods like fluorescence, mass spectrometry, electrochemical, and PCR-based detection techniques for microfluidic single cell analysis.

3.1 Fluorescence analysis

The fluorescence technique has attracted tremendous attention for use in analytical chemistry, cell biology and analytical instrument development. The fluorescence method possesses some obvious advantages that made it quite suitable for single cell analysis, for example, high sensitivity, simplicity, ease of integration, and the ability to provide spatially or temporally dynamic chemical information. There are many fluorescence detection methods such as fluorescence microscopy, laser induced fluorescence detection, flow cytometry, and so on. These methods have been greatly improved to be better compatible with the microfluidic chip, which could provide a more convenient and multi-parameter fluorescence detection of single-cell components compared to conventional methodologies. A new review has recently discussed the fluorescent probes, fluorescence labelling strategies, and the development of fluorescence detection equipment used for single-cell analysis on chip.83 In the following, we will critically introduce the recently developed new fluorescence detection methods on the microfluidic devices.

Fluorescence microscopy has long been used for intercellular biomolecule analysis, but the main limitation is that most small molecules lack intrinsic fluorescence and the biomedical activity will be greatly influenced after fluorescence labeling.84 Radionuclide labeling, which has been widely used in positron emission tomography (PET) scans for cancer diagnosis and monitoring, can overcome the drawbacks by keeping the structure of the labeled molecules.85 Surasi et al. recently employed the commonly used glucose analogue [18F]-fluorodeoxyglucose (FDG) as the radiotracer to trace its uptake in single cells using a fluorescence microscope on a droplet-based microfluidic system.85 Compared to conventional methods, the system enabled the measurement of the variability and modulation of FDG uptake at the single cell level. A few existing methods extract single cell information based on radioactive decays, but the process is stochastic, which will prevent the measurement of single cells in a high-throughput manner. With respect to this, a new concept for the conversion of random radioactive decays emitted from single cells into a permanent fluorescence signal was introduced.86 As shown in Fig. 4A, radiolabeled single cells and radiofluorogenic sensors were encapsulated in the droplets on a droplet-based microfluidic device. The principal mechanism of this assay is the reaction of the radiofluorogenic probe with reactive oxygen species like ROS produced by water radiolysis to generate a stable fluorescence signal. The fluorescence signal was proportional to the radioactivity level, which enabled quantitative measurement.


image file: c8an01186a-f4.tif
Fig. 4 (A) Workflow of the radiofluorogenic droplet assay. The mechanism of the radioassay is the conversion of radiofluorogenic probes to a fluorescence signal by reacting with ROS produced by water radiolysis. Reprinted with permission from M. E. Gallina, et al., Anal. Chem., 2017, 89, 6472–6481. Copyright 2017 American Chemical Society. (B) Schematic diagram of the multicolor fluorescence detection-based microfluidic device. Reprinted with permission from Q. Li, et al., Anal. Chem., 2016, 88, 8610–8616. Copyright 2016 American Chemical Society.

Single cell metabolomics study, which refers to the analysis of all intracellular small molecules, has a critical impact on the systems biology, stem cell research, and drug development.87 However, the large chemical diversity of metabolites, their inability to amplify and the fast metabolic turnover rates made them difficult to detect at the single-cell level. Microchip electrophoresis with laser-induced fluorescence detection (LIFD) is a good choice for metabolite analysis at the single cell level. However, most of the LIFD devices built with one-laser excitation and one-color fluorescence collection cannot fulfill the requirement of simultaneous detection of numerous metabolites. For this purpose, a new multicolor fluorescence detection setup was built by Tang's group, which used two lasers for simultaneous visible-light and near-infrared excitation and three optics channels for the green, red, and near-infrared fluorescence collection (Fig. 4B).88 In order to prove the potential of the established multicolor fluorescence detection system, intercellular small-molecule metabolites including hydrogen peroxide (H2O2), glutathione (GSH), and cysteine (Cys) were successfully separated and detected with high specificity and sensitivity at the single cell level.

3.2 Mass spectrometric analysis

Recently, mass spectrometry-based single cell analysis techniques have attracted increasing attention due to their label-free ability to detect unknown molecules. Combining microfluidic chip technology with mass spectrometry can simultaneously achieve the selectivity and sensitivity of single-cell analysis, which cannot be achieved by other technologies.89,90 In the following, we will discuss different ionization methods used in MS for single cell analysis and their developments in the past three years. To effectively analyze macromolecular compounds and provide an effective interface between liquid chromatography and mass spectrometry, electrospray ionization (ESI) proposed in the late 1980s provided an almost perfect solution to the above two requirements.91 A chip interface can now be integrated directly onto the electrospray ionization mass spectrometer using pressure-coupled or electroosmotic flow methods to introduce liquid into a mass spectrometer.92–94 Numerous researches have demonstrated that droplet-based microfluidics are powerful tools for high-throughput and multiple analysis of single cells. However, their combination with mass spectrometry has some difficulties. One major difficulty comes from the large difference between cell and droplet volume. The volume of typical mammalian cells is about 1 pL, but the volume of droplets generated from droplet-based microfluidics ranged from nanoliters to microliters, which will largely dilute the cellular components and cannot achieve mass spectrometry detection sensitivity. In addition, matrix effects, such as the culture medium or the intercellular matrix like proteins or small molecular metabolites, could not be ignored in single cell MS analysis.87 To address this problem, Zhang et al. developed a droplet-based extraction capillary to combine with ESI MS for cellular metabolite detection.95 A pulled glass capillary's tip containing extraction solvent was placed close to the surface of a single cell by a three-dimensional manipulator to extract cellular components. Besides, the glass capillary was linked to a syringe so that the extraction solvent can extrude or suck back easily. The interference of culture medium or other cellular components could be largely reduced after extraction procedures, and different desired metabolites could be easily extracted by using specific extraction solvents.

Matrix-assisted laser desorption mass spectrometry (MALDI MS) is a simple and effective soft ionization technique, which is widely used for the analysis of natural/synthetic biomolecules. It is a powerful tool for the analysis of biomolecules including proteins, carbohydrates, lipids, and polymers.96 The principle of MALDI is to use a laser to irradiate a co-crystallized film formed by the sample and the matrix. During the ionization process, the matrix absorbs laser energy and transfers part of its charge to analyte molecules to ionize them.97 An outstanding feature of MALDI MS is the ability to provide both chemical and spatial information inside single cells. Direct analysis of cellular biomolecules at the single-cell level has been demonstrated, such as a single yeast cell98 and a mammalian cell.99 However, the cells should be manually selected, which led to extremely low throughput. In order to realize high-throughput single-cell analysis by MALDI MS, our group developed a microwell-array-based microfluidic chip to combine with MALDI MS for automatic and high-throughput single cell phospholipid analysis.100 As shown in Fig. 5A, a cell array was formed on an indium tin oxide (ITO)-coated glass slide using a high-density PDMS microwell array. After matrix deposition, MALDI-MS imaging analysis at the single cells level could be automatically performed in high-throughput mode by setting an appropriate step size of the sample stage. Inspired by our work, Wu's group pioneered the use of a facile single-cell patterning method integrated with time-of-flight secondary ion mass spectrometry (TOF-SIMS) for high throughput analysis of drug-induced phenotypic alterations at the single cell level.101 A single-cell array with high site occupancy (>90%) and a high single-cell resolution (>97%) was obtained by combining a micropatterning PDMS stencil film with centrifugation-assisted cell trapping. Compared to other MS methods, TOF-SIMS has the ability to detect chemicals present on the surface with high sensitivity and spatial resolution. In addition, it holds significant advantages like no usage of a matrix and less susceptibility to contaminants. Due to the obvious merits of the combination system, cisplatin-induced phenotypic alternations of HeLa cells in the early stage of apoptosis were studied with high throughput.


image file: c8an01186a-f5.tif
Fig. 5 (A) Schematic illustration of single-cell capture and cell array detection by MALDI-MS. Reprinted with permission from W. Xie, et al., Anal. Chem., 2015, 87, 7052–7059. Copyright 2015 American Chemical Society. (B) Schematic diagram of the inkjet printing of cells onto the tungsten tip for MS analysis. Reprinted with permission from F. Chen, et al., Anal. Chem., 2016, 88, 4354–4360. Copyright 2016 American Chemical Society. (C) Schematic diagrams of chip design, droplet generation, and online droplet chip-ICPMS single cell analysis. Reprinted with permission from H. Wang, et al., Anal. Chem., 2017, 89, 4931–4938. Copyright 2017 American Chemical Society.

The recently developed probe electrospray ionization (PESI), with a sharp solid needle as an electrospray emitter, is also a powerful tool for single cell analysis. Lin's research group recently presented single-cell lipid profiling by combining drop-on-demand inkjet cell printing with PESI-MS.102 As shown in Fig. 5B, the droplet containing single cells precisely dripped onto the tungsten tip of the ESI needle. The tip enabled the direct spray of dripped droplets to a mass spectrometer by applying a high-voltage electric field so that lipid fingerprints of single cells could be obtained. Eight different types of single cells were differentiated according to cellular surface phospholipids using principal component analysis.

Numerous elements like carbon, nitrogen, phosphorus, iodine, and some metal ions like calcuium, iron, and chrome are essential for life.103 Mineral elements, even at a low concentration, play an important role in biological systems, such as the formation of biomolecules, regulation of gene expression, synthesis of small organic ligands and so on.104 So far, inductively coupled plasma mass spectrometry (ICP-MS) has been demonstrated to be a powerful technique for trace elemental analysis and has been applied for single cell analysis.105 The ion source of ICP-MS utilizes high temperature plasma to convert the atomic or molecular ion of an analytical sample into a charged ion.106,107 Compared to other inorganic mass spectrometry methods, ICP-MS has several advantages. Firstly, ICP-MS can be injected at atmospheric pressure and can be easily combined with other injection techniques. Secondly, the ICP-MS possesses a low detection limit, a fast analysis speed, a wide dynamic range, and a simple spectrum. Thirdly, isotopic analysis, single element and multi-element analyses, and morphological analysis of metal elements in organic matter can be performed. Finally, the initial ion energy is low and a simple mass analyzer can be used. Due to the above merits, ICP-MS is recognized as the most powerful trace and ultratrace multielement analysis technology.108 ICP-MS has been combined with a microfluidic based droplet generation system or on-chip solid phase microextraction for ultra-trace element analysis at the single cell level. Recently, Hu et al. integrated a magnetic solid phase microextraction (MSMPE) column on a chip to extract the release of Cd and Se in single cells after CdSe QDs treatment and then detected by ICP-MS.109 By optimizing the microextraction conditions, the limits of detection of Cd and Se could reach 2.2 and 21 ng L−1, respectively. After that, they developed a cross-channel droplet chip and online combined it with time-resolved ICP-MS via a miniaturized nebulization system for the quantification of Zn in single HepG2 cells (Fig. 5C).110 An essential requirement for time-resolved measurements is that cells should be spatially and temporally separated to ensure that each ICP-MS spike corresponds to one cell. In order to obtain the single cell encapsulation, the authors optimized the dimensions of the droplet generation channels and the sampling flow rate. Under the optimal conditions, the average diameter of the formed droplet is 25 μm with the droplet generation frequency of 3–6 × 106 droplets per minute, which could fulfill the requirement of high-throughput single cell analysis. Generally, high ratios of organic/aqueous phase are usually needed to generate stable and small droplets. However, the organic phase is always not compatible with MS detection. An important improvement of this article is that a microflow nebulizer and the addition of O2 gas were employed to avoid the usage of a large amount of organic solvent.

3.3 Electrochemical analysis

Electrochemical methods with miniaturized electrodes have a wide range of practical applications in biomedical sciences, such as the analysis of cell secretion in response to external simulation and the measurement of metabolites of individual living cells.111 Methylated DNA is considered a promising marker in the early diagnosis and prognosis of cancer, but the extremely low concentration makes it challenging to detect in clinical samples. Hong et al. reported an integrated microfluidic chip capable of preconcentrating methylated DNA using ICP and electrochemically detecting pre-concentrated DNA on a single chip.112 On the established platform, the original femtomole methylated DNA can be concentrated 100 times in a short time. Tian et al. incorporated an electrochemical DNA biosensor into a microfluidic paper-based analytical device for the detection of EGFR mutation in patients through a DNA hybridization reaction with high sensitivity.113 The negatively charged single stranded DNA without modified derivatization can be strongly absorbed onto the positively charged polypyrrole membrane modified gold electrode surface. The horseradish peroxidase finally catalyzes the redox reaction between H2O2 and methylene blue through differential pulse voltammetry, showing the rapid enhancement of current response.

Oxidative stress is one of the commonly analyzed parameters using electrochemical methods at the single cell level. Detection of superoxide anions released from skeletal muscle cells was recently reported by Ramin Banan Sadeghian.114 Nanoporous gold was used in an electrochemical biosensor with sensitivity increased 14 times, compared with nonporous gold, and they found a 3.7 nM limit of detection. Some amplification steps like using enzymatic labels can also be introduced into electrochemical analysis to improve the detection sensitivity. Safaei et al. integrated cell capture and the electrochemical detection method in one microfluidic system to detect very low levels of cancer cells in the whole blood.115 Compared with other electrochemical techniques, the sensitivity of this system increased remarkably due to the enzymatic nature of the assay. Glutathione can help maintain the function of the normal immune system, and has anti-oxidation and integrated detoxification. The sulfhydryl group on cysteine is easily combined with certain drugs and toxins (such as free radicals), and has a detoxification effect. Rapid and accurate detection of glutathione levels in individual cells is of great importance for diagnosis, prognosis and biomedical research on certain diseases.116 Shi et al. developed a new method based on microfluidic electrophoresis chemiluminescence to quantify glutathione in rat liver single cells, which is based on the strong sensitization of sulfhydryl compounds to the luminol-H2O2Cl system.117 Single cell injection, localization and membrane solubilization are rapidly performed in sequence on the microchip by a direct electric field force without any additional membrane solubilizing reagent.

Nanoparticles, widely used in many industrial applications, have attracted public attention regarding their potential toxicity. Many in vitro approaches were developed to study nanotoxicity, but the required cell labeling may interfere with nanoparticles, leading to false results. A chip-based approach and a label-free electrochemical impedance spectroscopy technique was developed by Li's research group to study the effects of nanomaterials on cell behaivor in real time.118 The electrode was integrated for electrochemical analysis and the DEP force helped single cell capture. They investigated the effects of CuO and TiO2 nanoparticles on PC12 cells, and the results showed that CuO nanoparticles induced higher toxicity than TiO2 nanoparticles.

Controlling the false positive rate is an important issue in clinical trials. Javanmard's research group integrated interdigitated multi-finger electrodes into microfluidic systems for minimizing false positives, especially for signals with a low signal-to-noise ratio.119 Compared with the commonly used single electrode pairs, multi-electrode sensors with higher signal-power were superior in distinguishing particles or cells from noise, showing the ability to characterize and identify unique single cells. Cans’ research group developed a new lithographic microfabrication technique to fabricate a movable thin film microelectrode array probe with 16 platinum band electrodes to record single cell exocytosis release events.120 Exocytosis release at multiple locations of a single cell surface could be simultaneously recorded by the 16 individual electrodes, resulting in nanometer scale resolution.

3.4 PCR-based analysis

Reverse transcriptase PCR (RT-PCR), a technique used to amplify the transcribed target RNA strands, has been widely used to reveal the heterogeneity of different cell types, such as stem cells and cancer cells. However, the presence of PCR inhibitors in low-quality biological samples can cause different degrees of reduction of the amplification efficiency, leading to false-negative results. To overcome the drawbacks, Mathies et al. developed a convenient and robust agarose droplet-based microfluidic approach to alleviate the effects of PCR inhibitors for single-cell and single molecule forensic short tandem repeat (STR) typing of samples.121 Primer barcoding beads and single cells were encapsulated in nanoliter agarose droplets using a microfluidic droplet generator. Cells were then lysed to release single-cell genomic DNA into the gel droplets. Before performing droplet PCR, the gel droplets were washed extensively to effectively remove PCR inhibitory molecules via diffusion from a porous gel. Compared to the conventional bulk PCR assay, the tolerance of urea, tannic acid, and humic acid in this method could be enhanced more than 10 times. In addition, for single cell PCR analysis on microfluidics, cells must be lysed to release RNA, but lysis agents and the lysate can inhibit the RT-PCR reaction. Some strategies can be taken to solve this problem, such as designing multiple microfluidic devices for cell lysis and PCR reagent addition separately. These methods inevitably increase operation complexity and reduce throughput. Recently, Abate et al. described a droplet microfluidic device integrating lysis and reagent addition for single cell RT-PCR analysis.122 Cells were encapsulated with alkaline lysis buffer in the droplets, where endogenous nucleases and inhibitory proteins could be denatured in a high pH buffer. The droplets containing lysed-cell were then merged with PCR droplets in the downstream channel for amplification. The multistep integration on one device provides a faster and more reliable method for single cell RT-PCR analysis.

MicroRNAs (miRNA), nonprotein coding and endogenous short RNAs, play vital roles in the regulation of biological process. Accurate quantification of miRNA in single cells will help us deeply understand their relationships with cell functions. Li et al. developed a ligation-based droplet digital PCR (ddPCR) to quantify miRNA in a single cell.123 In this ddPCR assay, the authors did not need to design a sophisticated reverse transcription for miRNA targets, but to simply design two target-specific oligonucleotide probes to complement the half-sequence of the target miRNA.

4. Applications for single cell analysis

4.1 Small molecule detection

Intercellular small molecules like metabolites, inorganic ions, and organic small molecules are widely involved in signal pathways, and play important roles in numerous physiological and pathological processes. For example, high glucose metabolism and lactate production is one of the outstanding features of cancer, and cancer cells are programmed to convert glucose to lactate through aerobic glycolysis.124 Abbyad et al. used a droplet microfluidic device to quantitative determine lactate release from many single cells by fluorescence based on the theory that lactate can be enzymatically converted to a fluorescence product.125 Reactive oxygen species (ROS) and reactive nitrogen species (RNS) are two important biomolecules in many oxidative-stress induced diseases and they always interact with each other.126,127 To fully understand their interaction and effects on the regulation of physiological events and oxidative-stress-induced disease, a microchip electrophoresis combined with laser-induced fluorescence was developed for the simultaneous detection of O2˙ and NO in single PC-12 cells before and after stimulation by 6-hydroxydopamine.77 This method employed a new consecutively gated injection for the microchip electrophoresis so that single cells could be consecutively injected with good reproducibility and without causing cell damage. H2O2, a particularly important representative of ROS, secreted from the cells, will influence cell migration, cellular communication, and immunity generation.128 A droplet-based microfluidic device combined with Au nanoclusters was developed for sensitive detection of secreted H2O2 from different types of single cells using an inverted fluorescence microscope (Fig. 6).129 The detection mechanism is based on the dramatic fluorescence changes of a horseradish peroxidase-Au nanocluster induced by secreted H2O2. This method showed an ultrahigh sensitivity of 200–400 attomoles H2O2 measured from a single cell. Various intercellular metal ions have structural and catalytic functions in protein and enzymes. For instance, in the nervous system, metal ions play important roles in neurotransmitter release and transmission, synaptogenesis, and synaptic transmission. Research studies have demonstrated that an excess or deficiency of essential metal ions can lead to neurodegenerative diseases.130 For example, the treatment of neuron-like cells with Aβ25–25 could induce variations in intercellular metal ion concentrations like Na+, K+, Ca2+ and Mg2+, which may be associated with Alzheimer's disease. Recently, an on-chip electrophoresis technique was employed to separate these four ions in Aβ25–25-treated single PC-12 cells, and these four metal ions were simultaneously detected by multicolor laser-induced fluorescence.131 This method provides a new route for multiple metal ions detection in single cells.
image file: c8an01186a-f6.tif
Fig. 6 Schematic illustration of detection of hydrogen peroxide in the single-cell encapsulated droplets in combination with HRP-AuNCs. Reprinted with permission from R. Shen, et al., Anal. Chem., 2018, 90, 4478–4484. Copyright 2018 American Chemical Society.

4.2 Protein analysis

Single cell-based quantification to reveal protein heterogeneity is extremely important in molecular biology, biomarker discovery, disease diagnostics, pathology and therapy.132 Quantitative analysis of specific proteins can greatly improve early diagnosis of diseases like cancer that originates from one cell or a small group of cells. Isobaric tags for relative and absolute quantitation (iTRAQ) are a widely used isobaric labeling method for quantitative proteomics by MS analysis. Recently, Ros et al. integrated a series of chambers and valves on a microfluidic platform to form a set of defined wells for the relative and absolute quantification of targeted proteins by MALDI MS/MS.133 In order to realize a quantification analysis, the iTRAQ labeling strategy was for the first time introduced into the microfluidic systems for protein labeling. All the necessary manipulation steps including protein digestion, labeling and matrix delivery could be carried out on the chip. They successfully detected the apoptosis-related protein Bcl-2 and quantitatively assessed the number of Bcl-2 molecules. From the results, they speculated that a microfluidic well dimension should be downscaled to a minimum surface area of ∼800 μm2 to fulfill the single cell detection sensitivity.

Invasion of malignant cells into the surrounding tissue is an important hallmark of cancer progression. Cells within a tumor have different invasive capacities. However, little is known about the relationship between the invasive phenotype with protein expression in individual cells. In order to address this problem, Kumar et al. introduced a live-cell imaging microdevice composed of polyacrylamide microchannels for the simultaneous measurement of invasive motility and protein expression in single human glioblastoma tumor-initiating cells (TICs).134 The procedures for western blotting analysis of protein levels in single cells including single cell separation, in situ lysis, and separation, could be achieved in one system. In this platform, they observed the relationship between motility and EphA2 expression in single TICs.

Numerous studies in the past few decades have shown that circulating tumor cells (CTCs) in peripheral blood samples are a fundamental prerequisite for metastasis. Researchers think that CTCs might be used as a marker to predict disease progression or disease diagnosis. However, clinical studies on CTCs show that there is significant functional heterogeneity in the CTC population. For example, only a small group of CTCs are liable to seed distant metastases, and a large fraction is prone to apoptosis.135 Due to this reason, clinical decisions usually benefit from the knowledge of multiple molecular profiles. Recently, Deng and Shi et al. presented a microchip-based approach for multi-detection of glucose uptake, intercellular functional proteins, and genetic mutations from individual CTCs.136 The microchip contains thousands of nanowells to isolate individual CTCs and their glucose uptake can be imaged using fluorescent glucose analogs. Moreover, intercellular signaling phosphoproteins were quantified using a miniaturized antibody barcode microarray. The individual cell nuclei could be finally retrieved for off-chip genome amplification and sequencing. This strategy broke the bottleneck of rare cell analysis with less than 20% cell loss.

Leukocytes, the essential cells in the immune system, play an important role in the protection of human body against foreign invaders. For example, metalloproteinases and disintegrins are important secretory products of individual leukocytes to regulate the inflammatory response and mediate host defense. However, cell heterogeneity of leukocytes has a significant effect on therapeutic outcomes of patients under inflammatory conditions and can be an indicator of disease progression.137 Characterization of the protease profile secreted from individual leukocytes can reveal the function of the immune system and personal physiological conditions for precision medicine. Chen et al. presented an integrated continuous-flow microfluidic system containing micropillars and a droplet generation system to purify, encapsulate and quantify proteases activity at the single leukocyte level.102 With this design, the native and phorbol 12-myristate 13-acetate (PMA)-simulated single leukocyte protease profiles were successfully measured at the single cell level, and it was found that PMA could elevate the average protease activity level and reduce cell-to-cell variation. Cytokines, mainly secreted by immune cells, play a critical role in infection, immune responses, inflammation, and disease development.138 The detection of cytokine proteins and related gene expression and transcriptional networks would be valuable for a better understanding of the immune responses. To address this purpose, co-detection of secreted cytokine proteins and genome-wide transcriptome from the same single cells was performed on a splittable single-cell microchip.139 Nanosized microchambers were designed on the PDMS layer to capture single cells with a multiplexed antibody array placed on top layer for the capture of secreted cytokine proteins. The proteins were then detected by ELISA after splitting the antibody array from the microchip, while the same single cells were then picked up by a 32G syringe for next generation sequencing of mRNAs (Fig. 7A). Through stimulating mouse macrophage cells with lipopolysaccharide, the proinflammatory factor TNFα and the anti-inflammatory factor IL10 were secreted by different cell types.


image file: c8an01186a-f7.tif
Fig. 7 (A) Scheme of detecting both secreted proteins and transcriptome from the same single cells. Reprinted with permission from J. George, et al., Anal. Chem., 2016, 88, 10309–10315. Copyright 2016 American Chemical Society. (B) Proteases produced within cell-encapsulated droplets cleave multi-color FRET substrates to yield multiple fluorescence signals. Reprinted with permission from E. X. Ng, et al., Biosens. Bioelectron., 2016, 81, 408–414. Copyright 2016 Elsevier.

Although there are many approaches for multiplex bioassays, most of them study average behavior across cell populations or have difficulty in real-time observation of actual enzymatic activity. A droplet microfluidic chip combined with multi-color fluorescence resonance energy transfer (FRET)-based protease-substrate reaction was developed by Chen et al. to simultaneously detect multiple specific protease activities at the single cell level (Fig. 7B).140 To prevent fluorescence cross-talk during the acquisition of signals, excitation and emission optical filters were added to the microscope filter cube. Six essential proteases activities were successfully detected using four enzyme–substrate reactions in three different types of cancer cells.

4.3 Multidrug resistance analysis

Multidrug resistance (MDR) is a big obstacle to tumor therapy, since cells exhibit resistance to distinct drugs. All the subpopulations with different genotypes and phenotypes have to be killed by therapy, because even small surviving cells may cause disease recurrence.141 In order to reveal some key information on the heterogeneity of cellular response to chemotherapy, always hidden in conventional analysis, high-throughput single cell MDR analysis has attracted an increasing attention in recent years. Besides, the isolation of drug resistant cells from a population of cells for further analysis has become one of the most important applications in the development of effective targeted therapeutic methods.142 A microfluidic device with a footprint of 20 × 28 mm2 enabling the simultaneous analysis of drugs response to more than 612 individual cells was presented by Dittrich's research group.143 Pneumatic doughnut-shaped valves were employed to trap and isolate cells, and help the cells in different sets of microchambers to be exposed to various drugs. Compared to the previously reported platforms, one main advantage of the platform is that one large channel housing all chambers can reduce the shear stress on the cells and avoid the clogging problem. The effects of three membranolytic anticancer peptides on MCF-7 cells were investigated by monitoring the efflux of calcein from single cells using fluorescence microscopy. The effect of the biomechanical heterogeneity of tumor cells on drug resistance was also investigated on an integrated microfluidic device.144 Single-cell arrays could be easily obtained by multi-obstacle architecture-like filter matrices, and a microvalve system was employed for controlled reagents and waste delivery. Vincristine resistance of single glioblastoma cells with different biomechanical properties was investigated on the platform. The throughput of the same single-cell analysis is also an important consideration in the MDR investigation. Recently, a single-step microfluidic dilution system, in which the gradient profile was only determined by the lengths of the distribution microchannel, was developed to stimulate cells cultured in the downstream culture chamber for high-throughput single-cell MDR analysis.145 MDR analysis in three different ways, namely time-dependent drug efflux kinetics, measurement of the secreted MDR-associated protein in the apical vacuoles of hepatic polarized HepG2 cells, and high throughput screening of the chemo-sensitizing effect and cytotoxicity of MDR modulators, could be flexibly performed.

4.4 Single cell sequencing with droplet microfluidics

As introduced above, microdroplets have significant advantages in high throughput applications due to their ability to produce droplets with kHz production frequency. Droplet-based microfluidics for single-cell sequencing has made some important progress over the last three years. A part or the whole genome of single cells can be measured in a sequencing device. To match the measured sequences with individual cells, different barcoding techniques were developed on droplet microfluidic devices.

Weitz and Kirschner et al. used droplet microfluidics to encapsulate the mixture of cells, lysis buffer, the ssDNA primer barcoding hydrogel microsphere and reverse-transcription reagents for subsequent analysis by next-generation sequencing.146 The method had the advantage of a low noise profile but faced the limitations of only 7% mRNA capture efficiency and a random barcoding strategy. Macosko and McCarroll et al. developed a drop-seq approach to analyse thousands of single-cell genomics simultaneously by separating, barcoding and sequencing of individual cells in nanoliter-sized aqueous droplets.147 With the improvement of throughput and correction of single cell sequencing, Klein and Mazutis et al. developed a droplet microfluidic technology, called in Drops, possessing the capability to index individual cells at a rate of >15[thin space (1/6-em)]000 cells per h.148 Cells and hydrogel beads with barcoding cDNA primers were first encapsulated in nanoliter droplets. Cells are then lysed and mRNA is captured by the released barcoding cDNA primers and reverse transcription. One benefit of this method is that each droplet carries primers encoding only one predefined barcode, which allows correction of PCR amplification bias during data analysis. Later, Abate's research group used droplet microfluidics to isolate single cells, lysis, fragment the genomes and barcode them, and finally the pooled DNA was sequenced on an Illumina MiSeq.149 Their method showed an ultra-high throughput that more than 50[thin space (1/6-em)]000 cells could be sequenced in a few hours.

5. Conclusions and future outlook

Significant improvements have been made in microfluidic technology for single cell analysis over the past three years. Microfluidics has become increasingly integrated, and shows high sensitivity and high throughput. Many significant applications have been reported in the fields of cell excretion detection, protein and multidrug resistance analysis using microfluidic systems. The trends of their applications have been changed from single target to multi-target detection. More research is currently focusing on molecular biology assays to explore the relationship between altered molecular information and clinical assessment, which indicates a significant research direction in the future. Although some advances have been made in microfluidic-based single cell analysis, some improvements are still needed like a more efficient single-cell manipulation ability and a detection instrument with faster and higher sensitivity. Because a large number of single cells need to be analyzed, output abilities as well as computational methods with powerful data processing capability need to be simultaneously improved. Due to the complexity of the cellular process, two or more methods can be combined together to detect more intercellular molecules. These challenges and also the new opportunities will stimulate the development of single cell analysis related manipulation, detection and multicomponent analysis, which will finally promote a more accurate and comprehensive understanding of the function and regulation mechanism of intercellular molecules in biological processes. These new methods will be of great importance for research and development in biology and medicine fields.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

The authors would like to thank Ziyi Pan of the Department of Chemistry at Stanford University for her assistance in editing the manuscript. Financial support from the National Natural Science Foundation of China (no. 21675096 and 21475073) and Youth Scientific Research Funds from Graduate School at Shenzhen, Tsinghua University (no. QN20160002) is highly appreciated.

References

  1. J. L. Spudich and D. E. Koshland, Nature, 1976, 262, 467–471 CrossRef CAS PubMed.
  2. F. Buettner, K. N. Natarajan, F. P. Casale, V. Proserpio, A. Scialdone, F. J. Theis, S. A. Teichmann, J. C. Marioni and O. Stegle, Nat. Biotechnol., 2015, 33, 155–160 CrossRef CAS PubMed.
  3. J. P. Junker and A. van Oudenaarden, Cell, 2014, 157, 8–11 CrossRef CAS PubMed.
  4. A. W. Hamburger and S. E. Salmon, Science, 1977, 197, 461–463 CrossRef CAS PubMed.
  5. D. Bonnet and J. E. Dick, Nat. Med., 1997, 3, 730–737 CrossRef CAS PubMed.
  6. C. E. Meacham and S. J. Morrison, Nature, 2013, 501, 328–337 CrossRef CAS PubMed.
  7. S. Lindstrom and H. Andersson-Svahn, Lab Chip, 2010, 10, 3363–3372 RSC.
  8. P. O. Krutzik, J. M. Crane, M. R. Clutter and G. P. Nolan, Nat. Chem. Biol., 2008, 4, 132–142 CrossRef CAS PubMed.
  9. P. O. Krutzik and G. P. Nolan, Nat. Methods, 2006, 3, 361–368 CrossRef CAS PubMed.
  10. J. El-Ali, P. K. Sorger and K. F. Jensen, Nature, 2006, 442, 403–411 CrossRef CAS PubMed.
  11. G. M. Whitesides, Nature, 2006, 442, 368–373 CrossRef CAS PubMed.
  12. X. He, Q. Chen, Y. Zhang and J.-M. Lin, Trends Anal. Chem., 2014, 53, 84–97 CrossRef CAS.
  13. J. Wu, Q. Chen, W. Liu, Z. He and J.-M. Lin, Trends Anal. Chem., 2017, 87, 19–31 CrossRef CAS.
  14. Q. Wu, D. Gao, J. Wei, F. Jin, W. Xie, Y. Jiang and H. Liu, Chem. Commun., 2014, 50, 2762–2764 RSC.
  15. M. C. Park, J. Y. Hur, H. S. Cho, S. H. Park and K. Y. Suh, Lab Chip, 2011, 11, 79–86 RSC.
  16. C. Liu, J. Liu, D. Gao, M. Ding and J.-M. Lin, Anal. Chem., 2010, 82, 9418–9424 CrossRef CAS PubMed.
  17. Z. Guan, S. Jia, Z. Zhu, M. Zhang and C. J. Yang, Anal. Chem., 2014, 86, 2789–2797 CrossRef CAS PubMed.
  18. M. M. Eric Brouzes, N. Savenelli, D. Marran, M. Twardowski, J. B. Hutchison, J. M. Rothberg, D. R. Link, N. Perrimon and M. L. Samuels, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 14195–14200 CrossRef PubMed.
  19. A. Huebner, M. Srisa-Art, D. Holt, C. Abell, F. Hollfelder, A. J. deMello and J. B. Edel, Chem. Commun., 2007, 28, 1218–1220 RSC.
  20. H. Wu, A. Wheeler and R. N. Zare, Proc. Natl. Acad. Sci. U. S. A., 2009, 101, 12809–12813 CrossRef PubMed.
  21. H. Zhang and K. K. Liu, J. R. Soc., Interface, 2008, 5, 671–690 CrossRef CAS PubMed.
  22. L. Weng, F. Ellett, J. Edd, K. H. K. Wong, K. Uygun, D. Irimia, S. L. Stott and M. Toner, Lab Chip, 2017, 17, 4077–4088 RSC.
  23. R.-J. Yang, L.-M. Fu and H.-H. Hou, Sens. Actuators, B, 2018, 266, 26–45 CrossRef CAS.
  24. R. Amann and B. M. Fuchs, Nat. Rev. Microbiol., 2008, 6, 339–348 CrossRef CAS PubMed.
  25. H. Wu, J. V. Volponi, A. E. Oliver, A. N. Parikh, B. A. Simmons and S. Singh, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 3809–3814 CrossRef CAS PubMed.
  26. D. K. Wood, D. M. Weingeist, S. N. Bhatia and B. P. Engelward, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 10008–10013 CrossRef CAS PubMed.
  27. C. Ma, R. Fan, H. Ahmad, Q. Shi, B. Comin-Anduix, T. Chodon, R. C. Koya, C. C. Liu, G. A. Kwong, C. G. Radu, A. Ribas and J. R. Heath, Nat. Med., 2011, 17, 738–743 CrossRef CAS PubMed.
  28. D. A. Lawson, N. R. Bhakta, K. Kessenbrock, K. D. Prummel, Y. Yu, K. Takai, A. Zhou, H. Eyob, S. Balakrishnan, C. Y. Wang, P. Yaswen, A. Goga and Z. Werb, Nature, 2015, 526, 131–135 CrossRef CAS PubMed.
  29. J. F. Zhong, Y. Chen, J. S. Marcus, A. Scherer, S. R. Quake, C. R. Taylor and L. P. Weiner, Lab Chip, 2008, 8, 68–74 RSC.
  30. F. Tang, C. Barbacioru, Y. Wang, E. Nordman, C. Lee, N. Xu, X. Wang, J. Bodeau, B. B. Tuch, A. Siddiqui, K. Lao and M. A. Surani, Nat. Methods, 2009, 6, 377–382 CrossRef CAS PubMed.
  31. F. Tang, K. Lao and M. A. Surani, Nat. Methods, 2011, 6, S6–S11 CrossRef PubMed.
  32. Y. Su, Q. Shi and W. Wei, Proteomics, 2017, 17, 1600267 CrossRef PubMed.
  33. W. W. Ning Gao, X. Zhang, W. Jin, X. Yin and Z. Fang, Anal. Chem., 2006, 78, 3213–3220 CrossRef PubMed.
  34. X. Gong, Y. Zhao, S. Cai, S. Fu, C. Yang, S. Zhang and X. Zhang, Anal. Chem., 2014, 86, 3809–3816 CrossRef CAS PubMed.
  35. M. K. Passarelli, C. F. Newman, P. S. Marshall, A. West, I. S. Gilmore, J. Bunch, M. R. Alexander and C. T. Dollery, Anal. Chem., 2015, 87, 6696–6702 CrossRef CAS PubMed.
  36. L. Lin, Q. Chen and J. Sun, Trends Anal. Chem., 2018, 99, 66–74 CrossRef CAS.
  37. L. Armbrecht and P. S. Dittrich, Anal. Chem., 2017, 89, 2–21 CrossRef CAS PubMed.
  38. T. W. Murphy, Q. Zhang, L. B. Naler, S. Ma and C. Lu, Analyst, 2017, 143, 60–80 RSC.
  39. S. Hosic, S. K. Murthy and A. N. Koppes, Anal. Chem., 2016, 88, 354–380 CrossRef CAS PubMed.
  40. A. Aharoni, G. Amitai, K. Bernath, S. Magdassi and D. S. Tawfik, Chem. Biol., 2005, 12, 1281–1289 CrossRef CAS PubMed.
  41. C. A. Suarez-Quian, S. R. Goldstein, T. Pohida, P. D. Smith, J. I. Peterson, E. Wellner, M. Ghany and R. F. Bonner, BioTechniques, 1999, 26, 328–335 CrossRef CAS PubMed.
  42. H. Yin and D. Marshall, Curr. Opin. Biotechnol., 2012, 23, 110–119 CrossRef CAS PubMed.
  43. Y. Wang, P. Shah, C. Phillips, C. E. Sims and N. L. Allbritton, Anal. Bioanal. Chem., 2012, 402, 1065–1072 CrossRef CAS PubMed.
  44. A. Nguyen, W. H. Khoo, I. Moran, P. I. Croucher and T. G. Phan, Front. Immunol., 2018, 9, 1553 CrossRef PubMed.
  45. J. F. Swennenhuis, A. G. Tibbe, M. Stevens, M. R. Katika, J. van Dalum, H. D. Tong, C. J. van Rijn and L. W. Terstappen, Lab Chip, 2015, 15, 3039–3046 RSC.
  46. L. Huang, Y. Chen, Y. Chen and H. Wu, Anal. Chem., 2015, 87, 12169–12176 CrossRef CAS PubMed.
  47. S. M. Park, J. Y. Lee, S. Hong, S. H. Lee, I. K. Dimov, H. Lee, S. Suh, Q. Pan, K. Li, A. M. Wu, S. M. Mumenthaler, P. Mallick and L. P. Lee, Lab Chip, 2016, 16, 3682–3688 RSC.
  48. A. Morimoto, T. Mogami, M. Watanabe, K. Iijima, Y. Akiyama, K. Katayama, T. Futami, N. Yamamoto, T. Sawada, F. Koizumi and Y. Koh, PLoS One, 2015, 10, e0130418 CrossRef PubMed.
  49. R. J. Jimenez-Valdes, R. Rodriguez-Moncayo, D. F. Cedillo-Alcantar and J. L. Garcia-Cordero, Anal. Chem., 2017, 89, 5210–5220 CrossRef CAS PubMed.
  50. E. Afrimzon, G. Botchkina, N. Zurgil, Y. Shafran, M. Sobolev, S. Moshkov, O. Ravid-Hermesh, I. Ojima and M. Deutsch, Lab Chip, 2016, 16, 1047–1062 RSC.
  51. T. Yang, D. Gao, F. Jin, Y. Jiang and H. Liu, Rapid Commun. Mass Spectrom., 2016, 30(Suppl 1), 73–79 CrossRef CAS PubMed.
  52. A. Korenaga, F. Chen, H. Li, K. Uchiyama and J. M. Lin, Talanta, 2017, 162, 474–478 CrossRef CAS PubMed.
  53. N. R. Schiele, D. T. Corr, Y. Huang, N. A. Raof, Y. Xie and D. B. Chrisey, Biofabrication, 2010, 2, 032001 CrossRef PubMed.
  54. J. J. Kim, K. W. Bong, E. Reategui, D. Irimia and P. S. Doyle, Nat. Mater., 2017, 16, 139–146 CrossRef CAS PubMed.
  55. W. Yang, H. Yu, G. Li, F. Wei, Y. Wang and L. Liu, Lab Chip, 2017, 17, 4243–4252 RSC.
  56. X. Jiang, D. A. Bruzewicz, A. P. Wong, M. Piel and G. M. Whitesides, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 975–978 CrossRef CAS PubMed.
  57. M. Thery, V. Racine, A. Pepin, M. Piel, Y. Chen, J. B. Sibarita and M. Bornens, Nat. Cell Biol., 2005, 7, 947–953 CrossRef CAS PubMed.
  58. C. A. Custodio, V. San Miguel-Arranz, R. A. Gropeanu, M. Gropeanu, M. Wirkner, R. L. Reis, J. F. Mano and A. del Campo, Langmuir, 2014, 30, 10066–10071 CrossRef CAS PubMed.
  59. M. G. Ahmed, M. F. Abate, Y. Song, Z. Zhu, F. Yan, Y. Xu, X. Wang, Q. Li and C. Yang, Angew. Chem., Int. Ed., 2017, 56, 10681–10685 CrossRef CAS PubMed.
  60. D. S. Shin, J. You, A. Rahimian, T. Vu, C. Siltanen, A. Ehsanipour, G. Stybayeva, J. Sutcliffe and A. Revzin, Angew. Chem., Int. Ed., 2014, 53, 8221–8224 CrossRef CAS PubMed.
  61. S. Okushima, T. Nisisako, T. Torii and T. Higuchi, Langmuir, 2004, 20, 9905–9908 CrossRef CAS PubMed.
  62. S. Y. Teh, R. Lin, L. H. Hung and A. P. Lee, Lab Chip, 2008, 8, 198–220 RSC.
  63. T. P. Lagus and J. F. Edd, J. Phys. D: Appl. Phys., 2013, 46, 114005 CrossRef.
  64. Z. Yu, L. Zhou, T. Zhang, R. Shen, C. Li, X. Fang, G. Griffiths and J. Liu, ACS Sens., 2017, 2, 626–634 CrossRef CAS PubMed.
  65. P. T. Kumar, K. Vriens, M. Cornaglia, M. Gijs, T. Kokalj, K. Thevissen, A. Geeraerd, B. P. Cammue, R. Puers and J. Lammertyn, Lab Chip, 2015, 15, 1852–1860 RSC.
  66. Y. Li, P. Chen, Y. Wang, S. Yan, X. Feng, W. Du, S. A. Koehler, U. Demirci and B. F. Liu, Adv. Mater., 2016, 28, 3543–3548 CrossRef CAS PubMed.
  67. X. Li, D. Zhang, H. Zhang, Z. Guan, Y. Song, R. Liu, Z. Zhu and C. Yang, Anal. Chem., 2018, 90, 2570–2577 CrossRef CAS PubMed.
  68. Y. Li, W. Zhang, J. Hu, Y. Wang, X. Feng, W. Du, M. Guo and B.-F. Liu, Adv. Funct. Mater., 2017, 27, 1606045 CrossRef.
  69. X. L. Guo, Y. Wei, Q. Lou, Y. Zhu and Q. Fang, Anal. Chem., 2018, 90, 5810–5817 CrossRef CAS PubMed.
  70. Z. Y. Li, M. Huang, X. K. Wang, Y. Zhu, J. S. Li, C. C. L. Wong and Q. Fang, Anal. Chem., 2018, 90, 5430–5438 CrossRef CAS PubMed.
  71. K. Zhang, M. Gao, Z. Chong, Y. Li, X. Han, R. Chen and L. Qin, Lab Chip, 2016, 16, 4742–4748 RSC.
  72. H. Chen, J. Sun, E. Wolvetang and J. Cooper-White, Lab Chip, 2015, 15, 1072–1083 RSC.
  73. M. Sauzade and E. Brouzes, Lab Chip, 2017, 17, 2186–2192 RSC.
  74. Q. Jin, M. Li, B. Polat, S. K. Paidi, A. Dai, A. Zhang, J. V. Pagaduan, I. Barman and D. H. Gracias, Angew. Chem., Int. Ed., 2017, 56, 3822–3826 CrossRef CAS PubMed.
  75. S. Casabella, P. Scully, N. Goddard and P. Gardner, Analyst, 2016, 141, 689–696 RSC.
  76. I. Y. Stetciura, A. Yashchenok, A. Masic, E. V. Lyubin, O. A. Inozemtseva, M. G. Drozdova, E. A. Markvichova, B. N. Khlebtsov, A. A. Fedyanin, G. B. Sukhorukov, D. A. Gorin and D. Volodkin, Analyst, 2015, 140, 4981–4986 RSC.
  77. L. Li, Q. Li, P. Chen, Z. Li, Z. Chen and B. Tang, Anal. Chem., 2016, 88, 930–936 CrossRef CAS PubMed.
  78. X. Li, B. Fan, S. Cao, D. Chen, X. Zhao, D. Men, W. Yue, J. Wang and J. Chen, Lab Chip, 2017, 17, 3129–3137 RSC.
  79. J. Liu, Y. Qiang, O. Alvarez and E. Du, Sens. Actuators, B, 2018, 255, 2392–2398 CrossRef CAS PubMed.
  80. Y. Zhao, K. Wang, D. Chen, B. Fan, Y. Xu, Y. Ye, J. Wang, J. Chen and C. Huang, Biosens. Bioelectron., 2018, 111, 138–143 CrossRef CAS PubMed.
  81. F. Guo, Z. Mao, Y. Chen, Z. Xie, J. P. Lata, P. Li, L. Ren, J. Liu, J. Yang, M. Dao, S. Sureshd and T. J. Huang, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 1522–1527 CrossRef CAS PubMed.
  82. D. J. Collins, B. Morahan, J. Garcia-Bustos, C. Doerig, M. Plebanski and A. Neild, Nat. Commun., 2015, 6, 8686 CrossRef CAS PubMed.
  83. Y. Fan, D. Dong, Q. Li, H. Si, H. Pei, L. Li and B. Tang, Lab Chip, 2018, 18, 1151–1173 RSC.
  84. L. Wei, F. Hu, Y. Shen, Z. Chen, Y. Yu, C. C. Lin, M. C. Wang and W. Min, Nat. Methods, 2014, 11, 410–412 CrossRef CAS PubMed.
  85. D. S. Surasi, P. Bhambhvani, J. A. Baldwin, S. E. Almodovar and J. P. O'Malley, J. Nucl. Med. Technol., 2014, 42, 5–13 CrossRef PubMed.
  86. M. E. Gallina, T. J. Kim, M. Shelor, J. Vasquez, A. Mongersun, M. Kim, S. K. Y. Tang, P. Abbyad and G. Pratx, Anal. Chem., 2017, 89, 6472–6481 CrossRef CAS PubMed.
  87. R. Zenobi, Science, 2013, 342, 1243259 CrossRef CAS PubMed.
  88. Q. Li, P. Chen, Y. Fan, X. Wang, K. Xu, L. Li and B. Tang, Anal. Chem., 2016, 88, 8610–8616 CrossRef CAS PubMed.
  89. A. Oedit, P. Vulto, R. Ramautar, P. W. Lindenburg and T. Hankemeier, Curr. Opin. Biotechnol., 2015, 31, 79–85 CrossRef CAS PubMed.
  90. L. Lin and J. M. Lin, J. Pharm. Anal., 2015, 5, 337–347 CrossRef PubMed.
  91. B. Domon and R. Aebersold, Science, 2006, 312, 212–217 CrossRef CAS PubMed.
  92. A. B. Theberge, F. Courtois, Y. Schaerli, M. Fischlechner, C. Abell, F. Hollfelder and W. T. Huck, Angew. Chem., Int. Ed., 2010, 49, 5846–5868 CrossRef CAS PubMed.
  93. R. Bandu, H. J. Mok and K. P. Kim, Mass Spectrom. Rev., 2018, 37, 107–138 CrossRef CAS PubMed.
  94. M. J. Pavlovich, B. Musselman and A. B. Hall, Mass Spectrom. Rev., 2018, 37, 171–187 CrossRef CAS PubMed.
  95. X. C. Zhang, Z. W. Wei, X. Y. Gong, X. Y. Si, Y. Y. Zhao, C. D. Yang, S. C. Zhang and X. R. Zhang, Sci. Rep., 2016, 6, 24730 CrossRef PubMed.
  96. Z. Lin and Z. Cai, Mass Spectrom. Rev., 2018, 1–16 Search PubMed.
  97. A. Tycova, V. Ledvina and K. Kleparnik, Electrophoresis, 2017, 38, 115–134 CrossRef CAS PubMed.
  98. A. J. Ibanez, S. R. Fagerer, A. M. Schmidt, P. L. Urban, K. Jefimovs, P. Geiger, R. Dechant, M. Heinemann and R. Zenobi, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 8790–8794 CrossRef CAS PubMed.
  99. S. Neupert, S. S. Rubakhin and J. V. Sweedler, Chem. Biol., 2012, 19, 1010–1019 CrossRef CAS PubMed.
  100. W. Xie, D. Gao, F. Jin, Y. Jiang and H. Liu, Anal. Chem., 2015, 87, 7052–7059 CrossRef CAS PubMed.
  101. L. Huang, Y. Chen, L. T. Weng, M. Leung, X. Xing, Z. Fan and H. Wu, Anal. Chem., 2016, 88, 12196–12203 CrossRef CAS PubMed.
  102. T. Jing, Z. Lai, L. Wu, J. Han, C. T. Lim and C. H. Chen, Anal. Chem., 2016, 88, 11750–11757 CrossRef CAS PubMed.
  103. J. S. Becker and N. Jakubowski, Chem. Soc. Rev., 2009, 38, 1969–1983 RSC.
  104. N. Unceta, M. Astorkia, Z. Abrego, A. Gomez-Caballero, M. A. Goicolea and R. J. Barrio, Talanta, 2016, 154, 255–262 CrossRef CAS PubMed.
  105. X. Wei, L. L. Hu, M. L. Chen, T. Yang and J. H. Wang, Anal. Chem., 2016, 88, 12437–12444 CrossRef CAS PubMed.
  106. S. E. Flores, A. S. Day and J. I. Keenan, Biometals, 2015, 28, 143–150 CrossRef CAS PubMed.
  107. W. Meng, Z. Zhenhua, M. Jin, W. Han, H. Man, Z. Fuxian, L. Yan, H. Bin, H. Zike and H. Qinxue, Biomaterials, 2015, 64, 78–87 CrossRef PubMed.
  108. R. F. S. Lee, S. Theiner, A. Meibom, G. Koellensperger, B. K. Keppler and P. J. Dyson, Metallomics, 2017, 9, 365–381 RSC.
  109. X. Yu, B. Chen, M. He, H. Wang and B. Hu, Talanta, 2018, 179, 279–284 CrossRef CAS PubMed.
  110. H. Wang, B. Chen, M. He and B. Hu, Anal. Chem., 2017, 89, 4931–4938 CrossRef CAS PubMed.
  111. A. Soldà, G. Valenti, M. Marcaccio, M. Giorgio, P. G. Pelicci, F. Paolucci and S. Rapino, ACS Sens., 2017, 2, 1310–1318 CrossRef PubMed.
  112. S. A. Hong, Y. J. Kim, S. J. Kim and S. Yang, Biosens. Bioelectron., 2018, 107, 103–110 CrossRef CAS PubMed.
  113. T. Tian, H. Liu, L. Li, J. Yu, S. Ge, X. Song and M. Yan, Sens. Actuators, B, 2017, 251, 440–445 CrossRef CAS.
  114. R. B. Sadeghian, S. Ostrovidov, J. Han, S. Salehi, B. Bahraminejad, H. Bae, M. Chen and A. Khademhosseini, ACS Sens., 2016, 1, 921–928 CrossRef CAS.
  115. T. S. Safaei, R. M. Mohamadi, E. H. Sargent and S. O. Kelley, ACS Appl. Mater. Interfaces, 2015, 7, 14165–14169 CrossRef CAS PubMed.
  116. S. Curello, C. Ceconi, A. Cargnoni, A. Cornacchiari, R. Ferrari and A. Albertini, Clin. Chem., 1987, 33, 1448–1449 CAS.
  117. M. Shi, Y. Huang, J. Zhao, S. Li, R. Liu and S. Zhao, Talanta, 2018, 179, 466–471 CrossRef CAS PubMed.
  118. P. Shah, X. Zhu, X. Zhang, J. He and C. Z. Li, ACS Appl. Mater. Interfaces, 2016, 8, 5804–5812 CrossRef CAS PubMed.
  119. P. Xie, X. Cao, Z. Lin, N. Talukder, S. Emaminejad and M. Javanmard, Sens. Actuators, B, 2017, 241, 672–680 CrossRef CAS.
  120. J. Wigstrom, J. Dunevall, N. Najafinobar, J. Lovric, J. Wang, A. G. Ewing and A. S. Cans, Anal. Chem., 2016, 88, 2080–2087 CrossRef PubMed.
  121. T. Geng and R. A. Mathies, Forensic Sci. Int.: Genet., 2015, 14, 203–209 CrossRef CAS PubMed.
  122. S. C. Kim, I. C. Clark, P. Shahi and A. R. Abate, Anal. Chem., 2018, 90, 1273–1279 CrossRef CAS PubMed.
  123. H. Tian, Y. Sun, C. Liu, X. Duan, W. Tang and Z. Li, Anal. Chem., 2016, 88, 11384–11389 CrossRef CAS PubMed.
  124. O. Warburg, Science, 1956, 123, 309–314 CrossRef CAS PubMed.
  125. A. Mongersun, I. Smeenk, G. Pratx, P. Asuri and P. Abbyad, Anal. Chem., 2016, 88, 3257–3263 CrossRef CAS PubMed.
  126. L. L. Qu, D. W. Li, L. X. Qin, J. Mu, J. S. Fossey and Y. T. Long, Anal. Chem., 2013, 85, 9549–9555 CrossRef CAS PubMed.
  127. Z. Chen, Q. Li, Q. Sun, H. Chen, X. Wang, N. Li, M. Yin, Y. Xie, H. Li and B. Tang, Anal. Chem., 2012, 84, 4687–4694 CrossRef CAS PubMed.
  128. K. Von Eynatten and G. Bauer, Int. J. Oncol., 2001, 18, 1169–1174 CAS.
  129. R. Shen, P. Liu, Y. Zhang, Z. Yu, X. Chen, L. Zhou, B. Nie, A. Zaczek, J. Chen and J. Liu, Anal. Chem., 2018, 90, 4478–4484 CrossRef CAS PubMed.
  130. P. J. Crouch and K. J. Barnham, Acc. Chem. Res., 2012, 45, 1604–1611 CrossRef CAS PubMed.
  131. L. Li, Y. Fan, Q. Li, R. Sheng, H. Si, J. Fang, L. Tong and B. Tang, Anal. Chem., 2017, 89, 4559–4565 CrossRef CAS PubMed.
  132. I. J. Majewski and R. Bernards, Nat. Med., 2011, 17, 304–312 CrossRef CAS PubMed.
  133. M. Yang, R. Nelson and A. Ros, Anal. Chem., 2016, 88, 6672–6679 CrossRef CAS PubMed.
  134. J. G. Lin, C. C. Kang, Y. Zhou, H. Huang, A. E. Herr and S. Kumar, Lab Chip, 2018, 18, 371–384 RSC.
  135. I. Baccelli, A. Schneeweiss, S. Riethdorf, A. Stenzinger, A. Schillert, V. Vogel, C. Klein, M. Saini, T. Bauerle, M. Wallwiener, T. Holland-Letz, T. Hofner, M. Sprick, M. Scharpff, F. Marme, H. P. Sinn, K. Pantel, W. Weichert and A. Trumpp, Nat. Biotechnol., 2013, 31, 539–544 CrossRef CAS PubMed.
  136. Y. Zhang, Y. Tang, S. Sun, Z. Wang, W. Wu, X. Zhao, D. M. Czajkowsky, Y. Li, J. Tian, L. Xu, W. Wei, Y. Deng and Q. Shi, Anal. Chem., 2015, 87, 9761–9768 CrossRef CAS PubMed.
  137. P. K. Chattopadhyay, T. M. Gierahn, M. Roederer and J. C. Love, Nat. Immunol., 2014, 15, 128–135 CrossRef CAS PubMed.
  138. G. Dranoff, Nat. Rev. Cancer, 2004, 4, 11–22 CrossRef CAS PubMed.
  139. J. George and J. Wang, Anal. Chem., 2016, 88, 10309–10315 CrossRef CAS PubMed.
  140. E. X. Ng, M. A. Miller, T. Jing and C. H. Chen, Biosens. Bioelectron., 2016, 81, 408–414 CrossRef CAS PubMed.
  141. R. A. Burrell, N. McGranahan, J. Bartek and C. Swanton, Nature, 2013, 501, 338–345 CrossRef CAS PubMed.
  142. W. M. Weaver, P. Tseng, A. Kunze, M. Masaeli, A. J. Chung, J. S. Dudani, H. Kittur, R. P. Kulkarni and D. Di Carlo, Curr. Opin. Biotechnol., 2014, 25, 114–123 CrossRef CAS PubMed.
  143. L. Armbrecht, G. Gabernet, F. Kurth, J. A. Hiss, G. Schneider and P. S. Dittrich, Lab Chip, 2017, 17, 2933–2940 RSC.
  144. L. Pang, W. Liu, C. Tian, J. Xu, T. Li, S. W. Chen and J. Wang, Lab Chip, 2016, 16, 4612–4620 RSC.
  145. Y. Li, D. Chen, Y. Zhang, C. Liu, P. Chen, Y. Wang, X. Feng, W. Du and B.-F. Liu, Sens. Actuators, B, 2016, 225, 563–571 CrossRef CAS.
  146. A. M. Klein, L. Mazutis, I. Akartuna, N. Tallapragada, A. Veres, V. Li, L. Peshkin, D. A. Weitz and M. W. Kirschner, Cell, 2015, 161, 1187–1201 CrossRef CAS PubMed.
  147. E. Z. Macosko, A. Basu, R. Satija, J. Nemesh, K. Shekhar, M. Goldman, I. Tirosh, A. R. Bialas, N. Kamitaki, E. M. Martersteck, J. J. Trombetta, D. A. Weitz, J. R. Sanes, A. K. Shalek, A. Regev and S. A. McCarroll, Cell, 2015, 161, 1202–1214 CrossRef CAS PubMed.
  148. R. Zilionis, J. Nainys, A. Veres, V. Savova, D. Zemmour, A. M. Klein and L. Mazutis, Nat. Protoc., 2017, 12, 44–73 CrossRef CAS PubMed.
  149. F. Lan, B. Demaree, N. Ahmed and A. R. Abate, Nat. Biotechnol., 2017, 35, 640–646 CrossRef CAS PubMed.

Footnote

These authors contributed equally to this work.

This journal is © The Royal Society of Chemistry 2019