Microfluidic Platforms for Single Cell Analysis: Applications in Cellular Manipulation and Optical Biosensing

Abstract

Cellular heterogeneity of any tissue or organ makes it challenging to identify and study the impact and the treatment of any disease. In this context, analysis of cells at an individual level becomes highly relevant for throwing light on the heterogeneous nature of cells. Single cell analysis can be used to gain insights into an overall view of any disease, thereby holding great applications in health diagnosis, disease identification, drug screening, and targeted delivery. Various conventional methods, such as flow cytometry, are used to isolate and study single cells. Still, these methods are narrower in scope due to certain limitations, including the associated processing/run times, the economy of reagents, and sample preparation. Microfluidics, an emerging technology, overcomes such limitations and is now being widely applied to develop tools for the isolation, analysis, and parallel manipulation of single cells. This review systematically compiles various microfluidic tools and techniques involved in single cell investigation. The review begins by highlighting the applications of microfluidics in single cell sorting and manipulation, followed by emphasizing microfluidic platforms for single cell analysis, with a specific focus on optical sensing-based detection in a high-throughput fashion, and ends with applications in cancer cell studies.

1. Introduction

Cells ranging from a few micrometers to tens of micrometers in size are the fundamental building blocks of life, which are dynamic and complex in nature. The advancement in the field of cell culture has proved that cellular heterogeneity and multimodal distributions exist within isogenic or clonal populations. In addition, the genetic information of the same phenotype of the cells may also be significantly different. Hence, the data obtained from any conventional cell culture study will be a collective averaged response of multiple cells of the population being studied, leading to a loss of crucial information in the background. In order to overcome such limitations, researchers are focusing on developing technologies for single cell analysis. Such systems are being developed to mainly address the issues posed due to cellular heterogeneity. Analyzing a single cell provides crucial information about the various cellular events affecting biological function, aiding the understanding of health and pathology at a deeper level [1,2]. Such an accommodative understanding also gives insights into various processes such as cellular proliferation, maturation, and homeostasis that help to recognize and identify diverse cell types. For instance, such heterogeneity can be seen in cancer, where cells vary in morphology, location, and characteristics. Assessment of circulating cancer cells at a single cell level is essential in developing patient-specific therapy.

An appropriately controlled environment is required to work with a single cell because excessive stress can disrupt and compromise the integrity of the cell. Intact single cells are crucial for various analyses, such as drug screening and patient status determination. Numerous technologies enable the assessment of individual cells and their molecular markers. Conventional methods for bulk cell population detection that are based on the measurement of overall magnetic properties and optical signals, such as magnetic resonance [3], high-throughput microscopy [4], and enzyme-linked immunosorbent assay (ELISA) [5], can identify cells from a large cell population. However, the increased demand for extremely precise and effective cell detection cannot be satisfied by these conventional methods because of their limited resolution and sensitivity due to bulk measurement [6,7,8,9,10]. The specificity and sensitivity of cell analysis are greatly improved by isolating and counting single cells. The most popular method for identifying cell characteristics and counting the precise number of cells is fluorescence-activated cell sorting (FACS). FACS functions by monitoring the fluorescence signals from a single cell [11]. Fluorescence tags related to the surface antigens of individual cells are used to identify each of them. Fluorescence light is produced when a cell is activated by a focused laser beam and propelled through the sensing zone. The wavelength and intensity, respectively, of the fluorescence signal, reveal the nature and density of the target receptors present on the cell surface. This technique has high throughput for the detection of multiple fluorescence tags. Furthermore, optical components such as an excitation light source, filters, detectors, and a sensitive cell-focusing mechanism must be used in order to improve the signal-to-noise (SNR) ratio. This makes the overall system heavy, complex, costly, and not user-friendly. Additionally, FACS demands a lot of cells and reagents and is susceptible to contamination when working with infectious samples [12]. Conventional molecular and cell biology benchtop instruments fall short of the requirements for precise and sensitive single cell measurements.

Microfluidic systems consist of microchannels in sub-millimeter-to-micrometer ranges, dealing with lesser sample volumes. The small size of microfluidic devices ensures laminar flow characteristics that make the fluid flow predictable and controllable. In general, these micro-platforms are created using polymers and gels, reducing the overall cost. Furthermore, microfluidic systems are advantageous over traditional methods in many ways, offering portability, rapidness, minimal sample volume requirement, cost-effectiveness, and parallelization. The application of microfluidics has now been extended from sample preparation to drug delivery and sensing applications. However, the number of developed techniques that have been commercialized is meagre for several reasons. One such reason may be the improvement required in the developed technologies to meet the rules and regulations framed by the various government organizations [13]. Despite this, these devices, referred to as lab-on-chip (LoC) devices or micro total analysis systems (micro-TAS), have gained attention as a crucial platform in biomedical research for diagnostic and bioanalytical applications [14]. Microfluidic systems for cell manipulation fall into different categories on the basis of different physical forces used, such as optical, magnetic, electrical, and mechanical. This review begins with an overview of the analytical and manipulation strategies for single cells, and mainly focuses on microfluidic platforms for single cell analysis, followed by their applications in optical biosensing.

2. Analysis of Single Cells

Microfluidics deals with small volumes of fluids, i.e., 10⁻⁹ to 10⁻¹⁸ liters in micrometer-sized channels. The structure and space of microfluidic devices provide opportunities for the precise manipulation of a single cell and the modification of the cellular microenvironment, as well as great flexibility in the design and manufacture of the devices. These also enable the evaluation of separation, and subsequent or parallel examination of single cell events [15,16,17].

2.1. Cellular Bioanalytics

It is well known that each cell residing in an identical environment can possess varied characteristic features. The ensemble detection methods manage only to provide averaged information data of the analyzed cell populations. Therefore, detection strategies for probing at the single cell level have flourished during current times. Such technologies usually comprise a sample preparation process flow for cell isolation, followed by a sensor platform. Several cell types are of critical importance in the areas of diagnostics and single cell-omics studies. For instance, pathological characterization of blood cells is performed for most disease diagnoses. Similarly, in the context of cancer, circulating tumor cells (CTCs) are considered a major and early indicator of tumor metastasis, which consists of a single cell or clusters of cells originating from tumor tissue [13,18]. Due to a CTC’s very low occurrence, its detection is immensely difficult, and thus strategies encompassing ultrasensitive and highly specific biosensing tools, and enrichment methodologies are currently being explored. The presence of molecular-level variations has been studied in different CTC populations and correlated with their metastatic potential. Performing such studies has been technically challenging using sophisticated technologies such as Next Generation Sequencing. Simpler versions of such systems customized for analyzing variations at the single cell level would help devise further developments in uncovering cell-to-cell heterogeneity. In the current time of pandemics, cellular variations have been studied extensively to understand several parameters, including host–pathogen interactions, host immune profiling, infection dynamics, and biomarker discovery [19,20]. These have helped to understand the infection mechanisms and eventually develop potential vaccine/treatment schemes. Under the scope of clinical diagnostics, utilization of single cell technologies has been rare and is still far from routine clinical analysis procedures. The lower utilization rate in the area of diagnostics is primarily associated with the high costs, tedious processes, and requirement of a trained workforce. Moreover, their translation as point-of-care (PoC) systems is steeply rising due to advances in conventional single cell analysis methods such as FACS [21], mass spectrometry [22], fluorescence, and confocal microscopic techniques, and the compatibility of these methods towards miniaturization [23]. The progress in the fields of sample preparation and cell isolation will be crucial in paving the way toward unravelling the cell complexities under normal and diseased conditions.

2.2. Sample Preparation

The accuracy of the single cell measurements largely depends on the sample preparation techniques involved. Sample preparation in molecular analysis of a single cell involves isolating cells from tissue, followed by purification, pre-concentration, and labelling of the samples. The first step of sample preparation is crucial in determining the subsequent processes of single cell analysis. In general, the biopsy is a well-explored way to obtain tissue to collect cells exhibiting the characteristics of their natural microenvironment. The extracellular matrix (ECM) and cell–cell junctions that bind cells together must be degraded to study individual cells from intact tissue. One of the crucial objectives of tissue separation is to obtain the maximum possible viable cells for further study. The traditional approach involves the enzymatic, chemical, and mechanical treatment of tissue to obtain individual cells. Pre-treatment of tissue with enzymes, including collagenase and trypsin, is used to degrade the proteins of ECM. Subsequently, chelating agents such as ethylenediamine tetraacetic acid (EDTA) have been used, and these bind to calcium ions leading to the disorganization of transmembrane proteins such as cadherins at the cell–cell interface. Following chemical treatment, an additional mechanical treatment involving frequent and gentle agitation was utilized to isolate cells in the suspension. For instance, Robin et al. demonstrated a technique to separate human myogenic cells from muscle tissue using collagenase activity; specifically, collagenase D and dispase II have been utilized in this study to obtain myogenic cells from a crushed muscle tissue sample [24]. Other dissociation enzymes used in the process of ECM degradation include hyaluronidase, elastase, and trypsin [25,26,27]. On the other hand, cell separation is not necessary for liquid biopsies such as blood sampling. White blood cells and circulating tumor cells (CTCs) can be separated from primary material such as blood using density gradient centrifugation [28]. Manual isolation of cells is efficient when studying a small number of cells. Various controllable parameters that can affect tissue separation are the concentration of chemicals, enzymes/chelators used, incubation time, and agitation mode. Nevertheless, these conventional typical mechanical and enzymatic tissue separation methods have certain limitations that make them less desirable [29]. Besides being labor intensive, the long protocols of these methods increase the chances of contamination, manual errors, and variances. Additionally, the expression of cell surface protein markers may decrease as a result of enzymatic digestion. Over time, methods for tissue dissociation have been developed and evolved to isolate cell samples with improved viability of delicate cells, while significantly eliminating the connective matrices that could obstruct the analysis of single cells. Laser capture microdissection (LCM) is an alternative method to enzymatic or digestion therapy [30]. However, the qualitative harvesting of cells using LCM is generally poor. Besides low throughput, this method also needs sample preservation and additional technical capabilities [31].

In order to overcome the limitations of conventional methods, researchers have developed modern tools such as microfluidic devices. For example, Hattersley et al. designed a microfluidic device for perfusion culture, analysis, and dissociation of intact rat liver tissue on the chip to better replicate the flux of in vivo tissue [32]. After treatment with ethylene glycol tetraacetic acid (EGTA), the immobilized tissue was perfused with Earl’s balanced salt solution (EBBS) to eliminate EGTA, which may interfere with collagenase inhibition followed by digestion due to dissolution in collagenase. Further, an ice-cold dispersal buffer was applied to the tissue at a high flow rate of 500 µL/min and, finally, the cells were collected at the output of the device into a centrifuge tube. Trypan blue assay results showed the cell viability after on-chip enzymatic separation was 78 ± 2.4%, similar to the results of the conventional dissociation techniques. This microfluidic method based on an enzymatic process shortens the time that tissue or cells are exposed to unsterile conditions, reducing the chances of contamination. However, the method could merely isolate 30,000 cells from a tissue of 4 mm³ after passing collagenase for 2 h [32]. Another problem for cell-based tumor tissue analysis, especially for clinical samples that are smaller and more difficult to collect, including those collected through needle aspiration biopsy, is poor cell recovery. A microfluidic device consisting of branching channels was created by Qui et al. to separate tumor tissue with a diameter smaller than or equal to 1 mm into individual cells [33]. Shear stress was created during the fluid flow due to the continuous constriction and expansion in every channel resulting in dissociation. Small HCT 116 colon cancer cell clusters, intact HCT 116 monolayers, and tumor spheroid made from LS 174T, HCT116, and NCl-H1650, respectively, cancer cell lines were used to test the effectiveness of the device. Non-enzymatic on-chip tissue dissociation demonstrated a considerable improvement in the single cell population from 61% to 95% while maintaining the same cell yields in intact cell monolayers compared to other conventional separation approaches utilizing trypsin-EDTA, pipetting, and vortexing. The mechanical dissociation technique employed in a microfluidic chip enabled non-enzymatic spheroid dissociation and increased cell production in less than 10 min among all studies [33]. Moreover, it offers high single cell yields, which are essential for small sample sizes. Additionally, the apparatus provides quick, non-enzymatic tissue separation, which is essential for determining endogenous molecular expression. When compared to other microfluidic applications, microfluidic tissue dissociation is still in its infancy, yet the benefits are already apparent. Undesirably, cells tend to collect or settle in the syringe and on the inner surface of the tubing while loading onto the chip, presenting technical difficulties for the microfluidic-based isolation methods. Researchers have either created conically tapered chip inlet regions, actively agitated the cell suspension in the syringe with a magnet, or used reagents with a similar density, such as Iodixanol [34], to avoid or reduce this problem. It is possible to alter channel geometries and flow rates to produce a “super-Poisson” encapsulation of cells, resulting in a significantly greater fraction of droplets with single cells [35,36]. Utilizing closed systems minimizes cell and/or tissue handling, lowering the risk of contamination. In laminar microfluidics, the geometry of the dissociation mechanisms is of a similar scale to biological samples, leading to better single cell yields, and flow conditions may be closely regulated, providing higher repeatability. Lower and more constrained clinical samples can be easily and successfully separated due to the smaller process volumes than that of the microfluidics feature. Moreover, the technology offers more single cells in output, which are essential for small sample sizes. Additionally, the system provides quick, non-enzymatic tissue separation, which is essential for determining endogenous molecular expression.

3. Microfluidics for Cellular Manipulation

Microfluidic cell analysis involves the trapping and separation of desired particles/cells from biological samples, holding applications in sorting, purification and enrichment of cells for cellular analysis, drug screening, and clinical and pathological diagnosis.

3.1. Trapping and Sorting of Single Cells

The ability of microfluidic systems to analyze samples at a single cell level enables them to be used in the sorting and trapping of cells. Microfluidic-based cell sorting techniques are divided into different systems, i.e., active and passive devices [37,38]. Cell sorters that require an external force generator to displace cells are referred to as active, while sorters that do not utilize an external force and are dependent on other ways, such as antibodies or geometry of channels, are known as passive devices. Different bioparticle manipulation forces (BMF) integrated with microfluidic technologies are dielectrophoresis, magnetophoresis, acoustophoresis, and optical tweezers, which affect cells/particles differently based on their varying physical features such as shape, size, and dielectric characteristics of the cells, as well as the culture media.

3.1.1. Micro-Trapping of Cells

The geometry of microfluidic trapping technology makes the isolation of single target cells possible and easy for researchers to understand cellular heterogeneity. The isolation performance of the microfluidic system can be affected by its physical structure. A single cell can be tracked with high precision by regulating the cell density or cell number. However, these physical features will significantly impair the cell culture and dynamic live cell imaging abilities because of geometric restrictions in the device’s spatial structure [16,39,40]. In 1997, the first approach for controlling a single cell on a microfluidic chip was reported [41]. Microfluidics-based single cell isolation is quickly becoming a crucial technique for recognizing and selecting target cells among a variety of cells present in biological fluids available for therapeutic use [42]. The most popular technique is to create microstructures on a microfluidic platform to trap or extract individual cells. Microstructures such as micro-trap and microwells resembling a single cell in size started years back and are now frequently used. The micro-trap is a type of self-regulating feature, as the change in dynamic flow surrounding the trap would considerably reduce the likelihood of further cells occupying it once a cell had already occupied the trap. T-shaped and V-shaped microstructures (Figure 1A) have been reported to analyze calcium ion concentration in different individual cells [43,44]. The two recent advancements in microfluidic arrays for single cell isolation are the dynamic single cell culture array and microfluidic array plate. These isolation devices can also be divided into two types on the basis of the driving force used. The first category covers the sedimentation of cells due to gravity, while the second one is attributed to the fluid drive [45]. Hydrodynamic and co-flow, along with micro-trap structures, can be used to trap the cells or particles.

A microfluidic device, which was fabricated for the trapping and manipulation of single cells, used actuation and sensing electrodes present in the trapping chambers connected to the main channel, inlets, and outlets, as shown in Figure 1B(i). The cells were first injected into the inlet of the device, which is then propelled to the main channel due to the applied external pressure and reaches the first actuation electrode, where negative dielectrophoresis pushes the cell into the trapping chamber. The particle encounters a vertical lifting force and lateral deflection as a result of the combined negative dielectrophoresis and hydrodynamic forces. Negative dielectrophoresis surpasses the drag force caused due to the hydrodynamic flow of the buffer and causes the cell to be retained in the trapping chamber Figure 1B(ii) [46]. Flow time, velocity, and cell density are critical variables for maximizing cell trapping efficiency without altering cell viability. The pumping of cells at a high flow rate speeds up the trapping phase and reduces the loading time. However, the high flow rate can lead to cell death as the cells are exposed to severe shear stress, which could break the integrity of the cell membrane. Another microfluidic platform with two inlets, two outlets, and a larger middle trapping chamber (where the cells are retained) based on micro-trap structures and co-flow has been reported for trapping individual cells. The device’s primary trapping chamber consisted of four rows of trapping arrays, each of which contained a pocket-like structure with three distinct geometries with a 6 ± 1 µm small opening (Figure 1C). Single cells were successfully captured in the trapping chambers at a flow rate of 1–3 µL/min, and the cytotoxic effects of different ethanol concentrations on hepatocytes were studied. The three different trap geometries that were used showed no differences, indicating that the shape of the trapping structures is not crucial if the gap of their aperture is smaller than the dimensions of the cells under study [47].

Magnetic labelling of cells with microstructures has also been explored to isolate and trap cells from a bulk sample. A trapping device has been developed and it comprises a silicon chip with microconductors and a microfluidic chip on the top of the device. After tagging with super-paramagnetic particles, CTCs were introduced into the device, and the cells were then retained in the trapping chamber at the base of the device. This retention happens because of the magnetic force applied to the cells, and the trapped cells remain in the trapping chamber even when the conductor is switched off, with the reason behind this being the micro-trap’s low fluid velocity. Additionally, neither the captured cell nor the empty micro-trap structures significantly affect the flow velocity of the fluid or sample over the cell [48]. These captured cells can then be used for further studies.

3.1.2. Droplet Encapsulation

Droplet encapsulation is an effective method to capture cells in an individual droplet. Materials can easily be internalized to form droplets by a test-tube-based emulsification method [49] via chemical or physical means such as ultrasonic-based oscillations [50]. Monodispersed droplets can be created with precise dimensional control using droplet microfluidics, a subtype of microfluidics. Each droplet might work independently as a bioreactor. The amazingly high throughput and precise fluidic controls that allow accurate mixing of fluids and compartmentalization of cells make droplet microfluidics one of the most appealing methods of droplet encapsulation [51]. Monodisperse water-in-oil nanospheres and microspheres, known as droplets, can be created using two immiscible fluids in microfluidic channels working with a volume range of a few femtolitres (fL) to nanoliters (nL) [52]. In addition to enabling high-throughput analysis with a small number of reagents, the characteristics of droplet microfluidics help to minimize solute–surface interactions. The target and constrained compartment concentration allow the droplet systems to achieve great detection sensitivity [53]. Additionally, the systems offer advantages such as simple attainment of quantifiable levels of analytes, the adjustable size of droplets to regulate the microenvironment, mimicking of relevant physiological conditions by encapsulating single cells with other chemicals, avoidance of cross-contamination among microreactors due to their restricted volume and distinct characteristics, and manipulation of droplets for purification, sorting, and other bioanalytical applications.

Microfluidic droplet creation is a straightforward, low-cost, passive, and efficient method for producing droplets with high efficiency. The utilization of a two-phase flow system in the specially constructed micropipe structure in the microfluidic chip can produce monodisperse droplets of various sizes [54,55]. Chabert and Viovy suggested a hydrodynamic method to encapsulate individual cells in picolitre droplets and the independent self-sorting of cells. Contrary to other systems, this approach precisely sorted the droplets as per the size of the encapsulated cell and effectively operated with heterogenous cell suspension, assuming that the cells do not undergo substantial cell–cell adhesion. Because of this hydrodynamic flow property, the authors were able to encapsulate a single malignant T cell from a suspension of whole blood, achieving a yield of more than 10,000 cells. Being a closed microchamber, the droplets make the addition of nutrients and outflow of metabolic waste difficult for single cell culture [56]. Droplets can be generated based on nozzle structures in the microfluidic device, microwell/microarray-based methods, and other methods. Droplet production based on designed nozzle structures such as T-junction, co-flowing, flow focusing, and DropChop geometry has advantages in the production of homogeneous droplets [57]. The insoluble continuous and dispersed phases are joined to form T-junction structures [58]. Droplets are created when the dispersed phase shears the continuous phase because of the pressure and shear stress. Because of the nozzle’s simplicity, it is considered a fundamental microfluidic design for the generation of droplets. In a coaxial focusing structure, droplets are created by aligning a capillary axis in the continuous phase fluid channel. The continuous phase presses the dispersed phase against the capillary tip to create the droplets.

A variation of the aforementioned approach used in microfluidics, known as flow focusing, is also used to encapsulate single cells. This focusing approach utilizes a vertical flow of continuous phases from opposite sides to shear the dispersed phase and creates droplets. The flow-focusing set-up entails crossing the flow of a cell medium and an oil-based medium at a focusing geometry and integrating it into planar microchannels [59]. Perfluorocarbon oils, being compatible with PDMS, immiscible with water, and generally transparent to facilitate optical detection, have been widely employed as the carrier fluid [60]. Droplet-based microfluidics was introduced by Hong et al. for screening drugs in cancer cell lines of primary tumors; in order to evaluate cell viability, single cells enclosed in droplets were imaged after 24 h of drug treatment [61]. Navi et al. demonstrated a technique utilizing flow-focusing microfluidic systems to form water-in-water droplets, where ferrofluid was used to achieve the diamagnetic separation between the empty droplets and droplets encapsulating the cell [62]. A microfluidic device for high-throughput generation and the sorting of microencapsulation with single cells was introduced by Nan et al. [63]. They used a sorting electrode to perform the selection of single cell encapsulation while the encapsulated cells were being transferred to the culture medium. Encapsulation with single cells triggered a fluorescent signal in an optical set-up with a counter-response to the sorting electrode with an output pulse, inducing dielectrophoretic force and hence forcing fluorescent cells to the collection channel, whereas encapsulations lacking fluorescent cells failed to trigger dielectrophoresis and flowed to the waste channel. The single cell encapsulation rate was 16% with a droplet diameter of 50 µm and a cell density of 3.05 × 10⁶ cells/mL, reaching more than 80% after cell sorting [63]. A droplet collection system was created by integrating fluorescence-activated sorting and a microfluidic droplet system that could perform a quantitative collection of droplets into microtubes without suffering sample loss. To achieve a continuous collection of single cell droplets, the integrated system used alternate sorting, distribution, and collection after droplet production in each branch channel. After being successfully trapped on a culture plate, the single cells that were encapsulated in droplets were then cultured to thoroughly assess the specificity and behavior of individual cells [64]. However, this approach of creating droplets for cell sensing is not preferable as it may damage the cell because of the stress caused by various non-physiological conditions and energy required to generate high voltage for driving droplets into channels.

3.1.3. Dielectrophoresis

There are different electric field-based techniques, such as electrophoresis, dielectrophoresis (DEP), electro-osmosis, and electro-rotation in microfluidics for cellular manipulation. The high and distinct dielectric characteristics of different cell types make DEP the most widely used technique for manipulating single cells. The uncharged cell or cell-like particles are dielectric in nature and start moving laterally in the presence of non-uniform electric fields due to polarization effects [65]. The particle becomes momentarily polarized as a result of the induced force due to a non-uniform field and migrates in the direction of either increasing or reducing field strength. The response of cells to an applied external field is caused by the ions and charged molecules present within the cell, which is also responsible for generating membrane potential. The appropriate combination of magnitude and frequency of electric potential applied across the electrodes account for cell trapping and are the major determining factors for the complete polarization of cells. The polarization of a particle depends on its physical characteristics, including permittivity and conductivity, as well as the characteristics of the medium [66]. The cells experience a force known as DEP, which also depends on their size and dielectric characteristics in reference to the surroundings. The electro-kinetic process has an application in the determination of viscoelastic properties of the cells. This DEP phenomenon can be employed in microfluidic systems, and the cells can be separated from a bulk sample by accurately regulating their electric field, flow control, and motion. Eukaryotic cells, bacterial cells, viruses, spores, and yeast have complex structures that do not alter the basic concept behind DEP but must be considered when formulating dipole moment and changing the DEP force [67]. DEP can be incorporated into microfluidics due to its excellent sensitivity, high analytical sensitivity, label-free nature, ease of manufacture, and compatibility [68,69].

Traveling wave DEP (twDEP) and interdigitated microelectrodes pattern were used by Cheng et al. to separate RBCs from other cells such as bacterial cells by focusing the cells into the channels followed by a twDEP application to separate the cells, as shown in Figure 2A [70]. A device was developed by Melvin et al., by combining DEP with electro-osmosis-based pumping in a microfluidic system, and they tested cell separation for various concentrations, particle sizes, and diluted electrolytes [71]. A microfluidic device based on continuous flow and multiple DEP forces was presented by Song et al. for sorting stem cells and their differentiation offspring, i.e., osteoblasts. This device manipulates and laterally deflects cells (Figure 2B) because of the multiple DEP forces and momentary electric field that switches on/off alternately. The collection efficiency and purity levels were 67% and 87% for osteoblasts and 92% and 84% for stem cells, respectively [72]. Insulators in microfluidic devices were also explored to sort cells by DEP. A DEP device based on an insulator was introduced to capture and detect yeast cells along with polystyrene particles. This system was able to successfully capture and enrich scarce/rare cells and particles with trapping efficiency greater than 99%, and the particles remain captured for more than 4 min (Figure 2C) [73]. The fabrication of these insulator-based systems is inexpensive and suitable since these devices consist of simpler components than other electrode-based systems and are generally made up of only one substrate. In addition to the capture, sorting, and manipulation of individual cells, DEP has also been exploited to detect the effect of drugs and drug inhibitors on single cancer cells. A microfluidic device integrated with DEP electrodes has been devised in which a microfilter structure filters and eliminates blood cells from cancer cells. DEP electrodes in the retention chamber of the device enabled quick and label-free isolation of individual prostate cancer cells, where drug accumulation and cell viability were also separately investigated in the presence of drugs and drug inhibitors [74]. DEP in integration with microfluidics has been used to detect the effect of drugs and drug inhibitors on single cells and identify the multi-drug resistant (MDR) positive and negative cells in an acute myeloid leukemia (AML) sample [75]. A microfluidic chip integrating DEP and impedance has been used to trap and count the MDR cells, respectively, from a sample containing wild-type cancer cells and MDR cancer cells based on differences in cytoplasmic conductivity [76]. This will aid in understanding heterogeneity in the inhibition of MDR activity in cancer samples.

3.1.4. Magnetophoresis

Electromagnets or permanent magnets can also be used to separate and sort individual cells for analysis. Since cells normally lack paramagnetic characteristics, surface tagging of cells with immunomagnetic beads is required. Because of the interactions between the cell surface antigen and the antibody-tagged nanomagnetic beads, they were attached to the surface of the cell in a specific manner. Cellular magnetic manipulation is convenient as it is easy to operate, because no heat is generated and contact with samples is also not required in this process. Moreover, the magnetic field and magnetism are not affected by the pH, ionic strength, and temperature of the surroundings [77].

Zhou et al. fabricated a PDMS-based microfluidic device exploiting magnetic forces to separate target particles from non-target particles in the samples. This device included two different parts—a separator, in the latter part where magnetic fractionation due to a magnetic field separates trapped target components from non-target components, and a cell incubator. This device trapped and isolated more than 90% of the target components and eliminated more than 99% of non-target components present in the sample. Integration of an incubator with the magnetism-based separator eliminates the need for manual off-chip incubation and makes automated separation and recovery of certain microparticles possible [78]. Hoshino et al. extracted cancer cells tagged with magnetic particles and sorted the cells by utilizing a microfluidic device based on a magnet array [79]. Intrinsic as well as induced magnetic properties of cells are used in microfluidics to separate cells. Red blood cells (RBCs) and white blood cells (WBCs) possess different intrinsic magnetic behavior. Deoxygenated RBCs are induced to possess paramagnetic behavior and so they move in the direction of the applied magnetic field, whereas oxygenated RBCs and WBCs are diamagnetic in nature and align randomly with reference to the magnetic field. A magnetic material made of nickel was used by Furlani et al. to separate WBCs and RBCs from blood plasma [80]. On being magnetized with permanent magnets, the deoxygenated RBCs were found to drift towards a microarray, while the WBCs travelled toward outlets. By employing the same strategy, Nam et al. used blood infected with malaria with a flow rate of 1.6 mL/min and separated RBCs from the sample (Figure 3A), achieving 99% efficiency [81]. Shields et al. successfully extracted magnetic CD4+ cells from blood with a 95% accuracy by applying micromagnets to microwells (Figure 3B) [82]. Magnetically activated cell sorting (MACS) applies magnetic microbeads/nanoparticles to label a target cell using antibodies for cell sorting. However, this technology has certain limitations, which are overcome by implementing the same in microfluidic devices that enable the generation of an intense magnetic field because of the rise in a magnetic field gradient across the cells [83]. Additionally, the microfluidic–MACS platform offers improved purity, better recovery rates, and utilizes fewer magnetic particles than MACS-based commercial systems.

3.1.5. Acoustophoresis

Acoustophoresis can be used as a contact-free and label-free approach to separate and sort cells depending on their biophysical features instead of biochemical characteristics. The phenomenon uses forces-of-sound radiation to manipulate particles of various materials ranging from micrometers to centimeters without making any direct contact. A variety of sample materials can be trapped using acoustophoresis-based or acoustic tweezers in different media. An acoustic tweezer was first used in 2012 to capture and control individual cells, microparticles, and complete organisms [84]. Bulk acoustic standing waves, traveling surface acoustic waves (TSAWs), and standing surface acoustic waves (SSAWs) are three different types of acoustic waves-based tweezers that have been employed in microfluidic devices. When ultrasound is used to excite microfluidic channels via the interdigital transducers (IDTs) into a resonance mode, where the applied wavelength is similar to the channel’s spatial dimensions, bulk acoustic standing waves are generated. The generated standing wave along the bottom of the microfluidic channel is responsible for positioning the cells or particles along the flow stream. Traveling surface acoustic waves use acoustic radiation force and acoustic streaming flow to actuate particles in microdevices [85]. Acoustic tweezers, which have been reported as utilizing focused acoustical vortices, are integrated with a microscope for the trapping and manipulation of selective individual cells from a cluster of cells. This device consists of metallic electrodes (i.e., IDTs) that generate spherical acoustic vortices, and a glass substrate placed between the microfluidic chamber and the transducer. Acoustic tweezers have been used for the trapping of cells with ~200 pN forces that showed no detrimental effect on the short- and long-term viability of cells [86].

Although acoustic-based methods have drawbacks such as low efficiency and sensitivity, these methods out-compete with other cell manipulation techniques for maintaining the phenotypic and genotypic characteristics of the cells after sorting. It is noteworthy that microfluidic-based acoustic systems offer solutions to overcome problems present in traditional acoustic methods. Acoustophoresis can be utilized to extract and manipulate CTCs due to differing acoustic forces that blood cells and CTCs experience based on their deformability, density, and size. Augustsson et al. created a device for the separation of prostate cancer cells and WBCs utilizing label-free and non-contact microfluidic acoustophoresis technology based on forces produced by ultrasonic resonance, as shown in Figure 4A. They differentiated cancer cells from WBCs and also examined the suitability of this technique using a blood sample from healthy volunteers after lysing RBCs and spiking with cancer cells. For non-fixed and viable tumor cells, they found that the purity ranged from 79.6% to 99.7%, with recovery ranging from 72.5% to 93.9%, while for paraformaldehyde fixed cancer cells, the purity was 97.4–98.4% and recovery 93.6–97.9% [87]. Guo et al. devised acoustic microfluidic tweezers based on three-dimensional trapping nodes that could trap single cells along the nodes [88]. A single beam acoustic tweezer based on microfluidics (Figure 4B(i)) was used by Hwang et al. to measure the mechanical characteristics of the target cells and the change in size of cells after trapping, as shown in Figure 4B(ii) [89]. Magnusson presented an automated and acoustic-based microfluidic device that can detect breast cancer cells and prostate cancer cells with equivalent purity and recovery yield. This device utilizes differences in the acoustic features of nucleated blood cells and cancer cells to align, as well as separate cancer cells from other cells, as presented in Figure 4C. A spiked blood sample volume of 5 mL that was processed indicated that WBCs obtained along with the cancer cell act as a contaminant and changing the voltage increased the purity of the cancer cells obtained in the outlet to 0.72%, having earlier been 0.11%, resulting in 62 ± 8.7% retrieval of cancer cells. Furthermore, the number of spiked samples did not affect the sorting capabilities for the investigation range of 10 to 2.5 × 10 ⁵ cells mL⁻¹ [90]. To eliminate the WBC contamination, another method was proposed by Anand et al., based on the two-step acoustophoresis approach. This method first separates cells on the basis of their varying acoustic features and mobility, and, subsequently, negative acoustophoresis and elastomeric particles (Eps) were used to remove the WBCs acting as contaminants from the obtained cancer cells, as shown in Figure 4D. Two 1 mL samples, (a) 10,000 DU145 cells in RBC-lysed whole blood (1:1 dilution) and (b) 1000 DU145 cells in 1.5 × 10⁶ WBCs, were assessed. The results showed a significant reduction in the number of WBCs, which were 459 ± 188-fold in sample (a) and 247 ± 156-fold in sample (b), corresponding to more than or equal to 99.5% efficiency. However, the recovery of viable cells for the (a) and (b) samples were 42% and 28%, respectively [91].

3.1.6. Optical Tweezers

In recent times, microfluidic devices with an optical system called opto-fluidics have gained great attention, as the technique is a contactless and contamination-free approach with applications in the extraction, sorting, and handling of individual cells. The target particle scatters light when a beam of light falls on it, leading to a change in momentum between the incident and scattered light photons, with an optical trapping force produced as a result. This is the basic mechanism involved behind particle manipulation in optofluidic technologies. Optical traps first illustrated by Ashkin [92] can investigate lengths as small as 0.1 nm, and can generate forces in atto-Newton and pico-Newton ranges. In an optical tweezer, cells are trapped by the opposing forces produced by two identical diverging laser beams. Mammalian cells and microbial cells have been sorted, transported, and assembled using optical tweezers [93]. In the case of CTCs, it is difficult to use optical force as the separation mechanism because CTCs and WBCs have indistinguishable optical constants. As a solution to this problem, the Yang group demonstrated a method by binding cancer cells (CC) to their homologous red blood cells (RBCs), as illustrated in Figure 5A. The authors used 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[folate(polyethyleneglycol)-2000] (ammonium salt) (DSPE-PEG-FA) to form folic acid-RBCs (FA-RBCs), which then bind to the folate receptors expressed on the surface of cancer cells. This resulted in the enhanced optical constant, i.e., refractive index units of tumor cells, which is significantly different from other cells and hence these CC-RBCs can be separated from other cells in an optofluidic platform by illuminating a laser beam onto the cell sample [94].

Liberale et al. created a device by combining micro-optical tweezers with microfluidic devices. This device was fabricated using two-photon photolithography and was able to optically trap and manipulate the single cells. The device was able to measure the Raman and fluorescence spectra of the individual cell, along with trapping single cells of diameter up to 20 µm [95]. Raman optical tweezers (ROT) is a technique incorporating Raman spectroscopy with optical tweezers which can be used for the manipulation and characterization of individual cells at the same time. Four different cells were segregated at a rate of approximately 90% using ROT. This also characterized and quantified the status of erythrocytes, leucocytes subtypes, and heterogeneous cancer cells. The complete set-up of this technique can be seen in Figure 5B [96]. In another work, the same group integrated RNA sequencing and RT-PCR of single cells with ROT (Figure 5C), with the technique used for transcriptomic profiling for the effective examination of drug sensitivity [97]. Despite being constricted to the manipulation of small numbers of cells, and the requirement of heavy and expensive instrumentation, it is one of the competing technologies for cellular manipulation in smaller sample volumes.

Figure 5. Optofluidic cell trapping (A) Optofluidic system showing modification of CTCs for their separation from RBCs based on differences in optical trapping constants (reproduced with permission from [94]). (B) Workflow of optofluidic system: (i) single cell is moved towards the slit (the stipe part of the hole) where ROT takes spectra, and (ii) isolated liquid from (i) is transferred to the circular zone to check for the presence of an isolated cell (adapted from [96]). (C) Schematic representation of the optofluidic system combining ROT for single cell Raman identification and single cell RNA sequencing (scRNA-seq) for the detection of drug sensitivity-related gene expression (reproduced with permission from [97]).

Figure 5. Optofluidic cell trapping (A) Optofluidic system showing modification of CTCs for their separation from RBCs based on differences in optical trapping constants (reproduced with permission from [94]). (B) Workflow of optofluidic system: (i) single cell is moved towards the slit (the stipe part of the hole) where ROT takes spectra, and (ii) isolated liquid from (i) is transferred to the circular zone to check for the presence of an isolated cell (adapted from [96]). (C) Schematic representation of the optofluidic system combining ROT for single cell Raman identification and single cell RNA sequencing (scRNA-seq) for the detection of drug sensitivity-related gene expression (reproduced with permission from [97]).

3.2. Optical Sensors for Single Cell Analysis

Optical sensing involves the measurement of variations in the optical parameters due to the corresponding biochemical interactions between the receptor and analyte. Broadly, the optical parameters are absorbance, reflectance, refractive index, phase shift, polarization, and energy of light. For single cell studies, organic dyes have been used conventionally, though currently the shift of use towards emerging technologies such as quantum dots and metal nanostructures has increased significantly. This shift is due to the issues related to photobleaching and the lower extinction capabilities of organic dyes. These emerging technologies usually involve plasmonic substrates which provide a plethora of interrogation techniques. Remarkable progress has been made in the directions towards interfacing microfluidics and optical detection systems (

Table 1. Single cell analysis using optical detection platforms.

Techniques	Target	Sensing Platform	Throughput/Assay Time	Applications/Conclusion	References
SPR	Leukemia cells (SHI-1)	Grating Coupled SPR	NA	Real-time cellular studies—adhesion and external stimuli at single cell level	[98]
	Living cells (lung and liver)	Cysteine-Derivative-Immobilized Au Films	5 SPR sensograms in 10 min	Multichannel SPR chip for single cell characterizations	[99]
SERS	Prostate cancer cells	Wheat germ agglutin-conjugated gold nanoparticles	~100 cells can be imaged and analyzed together using image processing	In-droplet cell-to-cell variability analysis in terms of size and number of glycan islands on cellular membrane	[100]
	Pathogens (E. coli, M. morganii, E. lactis, L. casei)	Silver nanoparticles	<5 min per sample	Pathogen identification and detection using spectral analysis	[101]
	E. coli cells	Silver and palladium bimetallic nanoparticles within porous silicon layers	10 sec	Cell detection, quantification, and monitoring of single pathogen type	[102]
Fluorometric	HeLa cells	Texas Red-tagged DNA stem–loops bound to silver nanowires	NA	Cellular temperature	[103]
	Human respiratory syncytial virus	Fluorescent aptamers targeted towards viral glycoproteins	6.5 μL/min	Counting and size determination of single viruses	[104]
	Influenza virus (IF)	2′-(4-Methylumbelliferyl)-α-d-N-acetylneuraminic acid, a fluorogenic substrate which emits fluorescence in the presence of neuraminidase present on the viral surface.	NA	Mobile imaging platform for virus detection and counting	[105]
Absorbance	Blood cells	Optofluidic platform monitoring cells at 405 nm	2000 cells/min	Differentiating between malaria-infected and healthy cells	[106]
	E. coli cells	Optofluidic platform monitoring cells at 280 and 600 nm	2100 spectra/sec	In-droplet analysis by UV-visible spectra acquisition and droplet sorting	[107]

SPR: surface plasmon resonance, SERS: surface-enhanced Raman spectroscopy, Au: gold, NA: Not available, UV: ultraviolet.

) for single cell analysis. Some of the primarily used modalities of optical detection are discussed below.

3.2.1. Surface Plasmon Resonance (SPR) Spectroscopy and Imaging

SPR sensing strategies have been conventionally used for studying biomolecular interactions in a real-time and label-free manner. SPR involves collective oscillations of free electrons induced due to incident electromagnetic waves at the metal–dielectric interface. At a specific angle (called the resonant angle) of the incident light, the intensity of the reflected light would be minimal due to the resonance condition. Slight refractive index variations in the vicinity of the surface would result in measurable changes in the resonant angle. Therefore, it has been an attractive method for probing molecular binding or interacting events at the metal–dielectric interface.

As SPR systems work in the near field, sensing is mainly dependent on the interaction between the bioreceptors and the surface antigens of the cells. Several kinetic studies at the single cell level have been uncovered through these techniques for monitoring the state of cells of interest [108]. These studies are relevant for uncovering the effects of antibiotics and therapeutic drugs. In the current age of antimicrobial resistance, a molecular understanding of cells would expedite screening and development of effective alternatives to classical antibiotics. Additionally, the detection of specific surface antigens, particularly those found on pathogenic and infected cells, can be translated to diagnostic applications.

Recent advances in optics have made SPR imaging (SPRi) a prominent detection technique, having the benefits of SPR spectroscopy plus the ability to detect spatial details. In summary, it is capable of imaging the variations in refractive indices at an interface. The instrumentation of SPRi systems was inspired by SPR systems. It involved a Kretschmann-like configuration of thin metal film coated on a glass prism. Charge-coupled device detectors are used for image acquisition to achieve spatial information by obtaining reflected light intensities from different sensing sites on the metal film or sensing layer. This has opened avenues for visualizing the entire working area and utilizing it for multiplexing applications. SPRi has been used for mapping cell–substrate adhesion and cellular surface movements in the case of cells. The cellular movement has been effectively utilized to study the effects of external factors and stimulants on cell dynamics, including cell growth, differentiation, detachment, and mobility [109,110]. In the context of clinical diagnostics, routine blood typing has been multiplexed and accelerated by immunoassays detecting RBC antigens [111,112]. Further, interfacing of SPRi with an inverted microscope enhances the technique’s spatial resolution and enables it to image events down to nanometer ranges. This enhancement becomes relevant in imaging, size, and mass determination of a single virus or bacteria. Due to their nano-size and low dielectric constants, these pathogens are difficult to detect with other modes [113].

Although SPR has shown impressive potential for the detection of cells, it faces some significant challenges in order to reduce issues associated with detection limits and image processing. Works on image processing, reconstruction techniques and response patterns are being explored currently for achieving high identification accuracy [99,114]. Further developments towards SPR integrated microfluidic platforms with compatible statistical tools could be very useful for the characterization of cells.

3.2.2. Surface-Enhanced Raman Scattering (SERS)

Surface-enhanced Raman scattering (SERS) is an important technique which provides spectral fingerprints that are excellent for the identification, and probing biochemical profile, of the analytes. The SERS substrate plays a big role in enhancing the scattering of light by the molecules upon their adsorption to the corrugated metallic surface. These corrugated or roughened metallic surfaces, typically made of gold or silver, bind strongly with thiols, amines, thioethers, etc., present in the biological molecules, making them ideal for detection. The corrugated features concentrate and enhance the incident electromagnetic radiation to interact with the nearby analyte molecules. The amalgamation of microfluidics with SERS has provided significant strides for the development of Raman-activated cell sorting platforms capable of identifying and sorting single cells from complex media [115,116]. Recently, Wang et al. demonstrated the applicability of these amalgamated systems for high-throughput analysis of in vivo enzyme function and sorting of cells at up to 120 cells per minute [117]. Willner et al. showed one of the earliest integrations of microfluidics and SERS by demonstrating the heterogeneity of size and shape of glycan islands on the cellular membrane of single cells [100]. Similar strategies have been expanded for detecting cell secretions which becomes relevant in cases of disease diagnosis and treatment (Figure 6A) [118]. With the ability to record single cell Raman spectra, which essentially provide a unique collection of Raman peaks specific to the targeted single cell, it has immense potential to be used for diagnostic and screening purposes. For instance, the integration of modern statistical tools and pattern recognition techniques have been used for the detection of bacterial strains from SERS spectral datasets (Figure 6B) [119]. The road ahead in this area should be to focus more on understanding specific phenotypes and in vivo functions for monitoring cells. Additionally, optimization with machine learning algorithms would smooth the process further and provide opportunities to multiplex through spectral pattern analysis.

3.2.3. Absorbance-Based Detection

Absorbance-based screening methods have been leveraged extensively for conventional cell viability studies. The success of this method is due to its relatively simple nature and high efficiency in interrogating absorbance of molecular species over a broad spectrum of light wavelengths. Cellular analysis becomes even more practical with this mode due to high absorbance of nucleic acids and aromatic amino acids in the UV region (<300 nm). Biosensors employing a 280 nm light source for quantifying cell populations by exploiting the spectral nature of amino acids have been widely demonstrated [120].

The instrumentation system typically involves a pair of a light source and a spectrometer/photodetector, enabling the system to be used for PoC applications. Microfluidic interfacing has given a boost towards realizing high-throughput systems. In the work by Banoth et al., a handheld PoC absorbance-based fluidic analyzer was fabricated for the quantification of malaria-infected RBCs and healthy RBCs [106]. Absorbance at a wavelength of 405 nm was monitored and depending on the detector voltage output, the cells were differentiated at a rate of 2000 RBCs per minute (Figure 6C(i,ii)). Recently, Duncombe et al. developed a UV-visible spectroscopy-based in-droplet analysis and sorting system for the detection of cells by measuring optical density at 280 and 600 nm (Figure 6D(i,ii)). The characteristic absorbance peak was selected at 280 nm for proteins, while 600 nm was chosen for measuring the scattering of light by the cells [107]. The system could also be modulated for monitoring other molecular events due to its ability to combine spectral acquisition and the sorting of droplets of interest.

3.2.4. Fluorometry

Fluorescence microscopy is a widely used technique for visualizing cells and spatio-temporal monitoring of cell growth and morphological changes. Flow cytometry is such a fluorescence-based technique in which cells are made to flow past lasers for high-throughput quantification and sorting of cells. The scattered or fluorescence light photons are counted, which essentially indicate the number of cells present in the sample. Interestingly, the monitoring of single cell transcriptomics or mRNA expression levels using flow cytometry has made huge developments recently. Conventionally, flow cytometric analysis for mRNA was limited to abundantly expressed genes. Recent advances in isothermal in situ nucleic acid amplification techniques have overcome such limitations. Zhao et al. used rolling circle amplification for detecting the target mRNA, followed by DNA amplification (Figure 7A). Fluorescent probes were hybridized with the amplified products and eventually boosted the signal [121].

Fluorescence thermometry is also emerging as a powerful analysis tool for the measurement of temperatures at a single cell level with high accuracy and spatial resolution [122]. Usually, temperature-sensitive fluorophores that are sensitive to the local temperature fluctuations are employed and the spectral properties of the same are monitored for sensing the temperature in a label-free manner. For example, a silver nanowire decorated with thermo-sensitive fluorophore-tagged DNA loop was used for temperature sensing in the range of 20–50 °C with a sensitivity of 2.6%/°C and temperature resolution of 0.5 °C [103]. The silver nanowire acted as a fluorescence quencher while the DNA loop changed its structure that led to enhanced fluorescence, depending on the local temperature (Figure 7B(i,ii)). Similar sensing schemes were expanded towards monitoring and detection of other entities including H⁺ ions (pH) and small molecules such as ATP [123].

Figure 7. Fluorometry–based sensing Schematic representation of (A) monitoring of single cell gene expression levels using padlock probes which are designed specifically to detect target nucleic acids, which is then ligated by ligase enzyme, amplified through rolling circle amplification, and analyzed through flow cytometry. FSC: forward scatter, SSC: side scatter (reproduced with permission from [121]). (B) Single cell DNA thermometer functioned through conformational changes (i) and variations in fluorescence intensity (FI) (ii) at two temperatures (reproduced with permission from [103]). (C) (i) Microfluidic in-droplet analysis of antibiotic resistance detection using fluorescence sensing and (ii) microscopic image of the sensor chip (reproduced with permission from [124]). (D) Single cell counting and size determination using fluorescent aptamers attachment to targets (reproduced with permission from [104]). (E) Smartphone interfaced fluorescent imaging platform for digital counting of virus (reproduced with permission from [105]).

Figure 7. Fluorometry–based sensing Schematic representation of (A) monitoring of single cell gene expression levels using padlock probes which are designed specifically to detect target nucleic acids, which is then ligated by ligase enzyme, amplified through rolling circle amplification, and analyzed through flow cytometry. FSC: forward scatter, SSC: side scatter (reproduced with permission from [121]). (B) Single cell DNA thermometer functioned through conformational changes (i) and variations in fluorescence intensity (FI) (ii) at two temperatures (reproduced with permission from [103]). (C) (i) Microfluidic in-droplet analysis of antibiotic resistance detection using fluorescence sensing and (ii) microscopic image of the sensor chip (reproduced with permission from [124]). (D) Single cell counting and size determination using fluorescent aptamers attachment to targets (reproduced with permission from [104]). (E) Smartphone interfaced fluorescent imaging platform for digital counting of virus (reproduced with permission from [105]).

Nowadays, miniaturized optofluidic platforms have also been developed which provide on-chip sample preparation, manipulation, and quantification of cells [125]. For instance, Lyu et al. were able to demonstrate automated phenotypic microscopic quantification of antibiotic resistance at the single cell level through the droplet encapsulation of cells (Figure 7C(i,ii)). The droplets act as test chambers for reactions that involve the cells, antibiotics, and a cell viability fluorescence marker [124]. Developments towards digital detection platforms facilitating the detection of single viruses are gaining impetus for early diagnosis and biological security [126]. For example, single viruses were captured using specific fluorescent-tagged aptamers for detection, counting, and sizing by nanoparticle tracking analysis (Figure 7D) [104]. However, for translation to PoC devices with many such systems, further efforts are needed along the lines of interfacing. In this direction, Minagawa et al. produced a digital detection method for influenza virus counting using a fluorometric method, interfaced with a smartphone camera-based image readout (Figure 7E) [105].

3.3. Single Circulating Tumor Cell Detection

CTCs are considered a significant biomarker of cancer diagnosis and prognosis, though the challenges emerge typically in their detection. A milliliter of blood contains millions of white blood cells, billions of red blood cells, and only a few CTCs. Microfluidic enrichment strategies have been proposed harnessing the physical and biological properties of the CTCs, as discussed in the sample preparation section of this review. Presently, CellSearch, an FDA-certified system, is considered the gold standard for capturing and quantifying CTCs from approximately 7.5 mL of blood sample [127]. Nevertheless, this technology is only limited to CTCs expressing epithelial cell adhesion molecules. Alternative methods for detection and enumeration of CTCs are currently being explored. SERS nanoprobe was prepared to utilize tetrahedron DNA assembly to electrostatically interact with positively charged AuNPs to obtain a tetrahedron AuNP assembly (Figure 8A(i)) [128]. The nanoprobe utilized EpCAM aptamers which bind specifically to the surface antigens of CTCs. The nanoprobe-enabled SERS spectral analysis (Figure 8A(ii)) could detect 3–500 CTCs/mL under optimal conditions.

In vivo tracking of CTCs has also been proposed by Dutta et al. by using gold nanorods (GNR) as labels for optical coherence tomography (OCT) (Figure 8B) [129]. In addition, the Yang group demonstrated a volumetric bar chip capable of visual detection of CTCs at a single cell level (Figure 8C) [130]. The CTCs were captured by aptamer-conjugated nanoparticles and the oxygen released from the reaction of nanoparticles (acting as catalase) and hydrogen peroxide leads to the dye travelling a distance that depends on the CTC density. Depending on the distance travelled by the dye, the concentration of the CTC can then be determined. Apart from this, Zhu et al. devised a strategy utilizing the peroxidase activity of SWCNT in the presence of messenger DNA (mDNA sequences) [131]. Aptamer-conjugated Fe₃O₄–SiO₂ particles bound to DNA, upon interacting with CTCs, leads to the release of the mDNA from the aptamer nanoparticle complex (Fe₃O₄–SiO₂-Gel/P1/mDNA) onto the solution. The single-walled carbon nanotubes (SWCNTs) present in the solution in the presence of released mDNA help in colorimetric detection in the TMB-H₂O₂ system due to the peroxidase-like activity of SWCNTs (Figure 8D). Lee et al. have also shown remarkable progress towards on-chip molecular profiling of single CTCs from metastatic breast and pancreatic cancers’ gene expression analysis without any off-chip processing [132]. Overall, diagnostic platforms are developing rapidly towards the ultrasensitive detection of CTCs through high sensitivity and specificity. It would be ideal if work towards efficient CTC capture systems is further explored and subsequent analysis of CTC subpopulations using simpler PoC systems are achievable. Looking ahead, cell-free DNA has also emerged as an exciting new biomarker capable of unravelling the genetic landscape of a tumor [133]. Additionally, the minimal sample requirements and early-stage prevalence make cell-free DNA ideal for designing precise and personalized treatment and monitoring regimes. Technologies towards integrated microfluidic platforms, as have been demonstrated by Cheng et al. for fluorescence monitoring, are crucial for providing a robust screening performance that becomes indispensable for early-stage diagnostic development [134].

4. Conclusions

Recent reports in the cellular biotechnology domain have shown that single cell studies are crucial as they provide insights towards accurate and early diagnosis of pathological conditions. The primary step in systematically analyzing a single cell requires pre-processing or preparation of the sample to obtain single cell populations. The conventional methods for such sample pre-processing protocols are generally time-consuming, and they usually demand large volumes of reagents and sample. To overcome these shortcomings, several microfluidic technologies have been introduced to address the associated spatio-temporal limitations. Initially, the development of microfluidics was restricted to mainly physics and engineering applications, but now it is being implemented in various domains. One such domain is single cell analysis, the outline and details of which have been discussed in this review. Integrating optical sensing with microfluidics enables selective manipulation and detection of individual cells. Microstructures in the microfluidic device can trap and sort desired single cells without the help of any external support. Some techniques that employ external forces, i.e., droplets, dielectrophoresis, magnetophoresis, acoustophoresis, and optical tweezers, have also been integrated in the microfluidic devices, and have been well explored for the manipulation of single cells. Moreover, in droplet microfluidics, microdroplets with an inert carrier oil encapsulate individual cells which can be used for further manipulation, reducing the risk of sample contamination [135]. Optical detection is a potential contactless technique and is label-free, wherever required, that does not exert any shear stress on the cells, and its incorporation with microfluidics enhances the strengths of the system by surpassing important challenges of single cell manipulation. Optical detection techniques enable sensitive detection, and one of the techniques that can detect individual cells precisely. The progress made in the field of microfluidics-based single cell manipulation and detection so far has enabled researchers to delve deeper into understanding cellular heterogeneity in several diseased conditions.

High-throughput strategies for the analysis of millions of cells in a few minutes, which were reviewed in this article, have the potential to directly affect the conventional design of cell therapy procedures. However, there are certain limitations in the aforementioned approach which need to be addressed. Optical detection technologies require sophisticated equipment and a complex set-up to perform analytical procedures. Future work can be conducted in the direction towards improving and developing the portability of these systems. Furthermore, commercialization of the developed techniques requires proper liaison and communication between engineers, researchers, and industrial partners. Democratization of these technologies through commercialization would be a significant step towards the transformation of currently known technologies into a smart and portable single cell analyzer system.