Microfluidic devices constrain fluids in tiny channels, allowing for precise control of samples as small as a few nanolitres or picolitres. Microfluidics-based assays are highly multiplexed (they can allow multiple elements to be measured) and provide high resolution and sensitivity for detecting nucleic acids, proteins and small molecules – all while saving costs, since so little of the sample is used. The last decade has seen microfluidics become increasingly entwined with another technology: single-cell analysis.
Studies show that identical cells from the same tissue could have highly diverse genotypes or phenotypes. This cellular heterogeneity underlies the complexity of cellular processes and their physiological outcomes, such as the onset of disease. To study an individual cell, you first need to isolate it – and microfluidics is a handy solution here.
“A key challenge with single-cell analysis is you need to measure small numbers of molecules present within one cell,” says Dino Di Carlo, professor of bioengineering at the University of California, Los Angeles.
Here, microfluidics allows single cells to be boxed within sub-nanolitre volumes. “When a cell secretes something or is lysed and releases something, it gets concentrated at a relatively high level and you can detect it more easily,” he adds.
Microfluidic single-cell analysis can achieve singlecell resolution in detecting metabolites, ions or small molecules. This generates detailed data into how these molecules regulate cellular processes, which are insights that could advance novel therapies.
Microfluidic droplets
Researchers employ different microfluidic approaches to trap single cells. These include cell-sized microwells and micropatterning, a technique to etch surface patterns with sticky spots to bind cells. However, these methods are limited by poor capture efficiency and non-uniform cell distribution, respectively.
That’s why the focus of single-cell microfluidics research has shifted to droplet microfluidics in recent years. It involves capturing cells inside tiny droplets, with little risk of cross-contamination. These cells can then be isolated: varying the flow rates of the cell suspension and oil-based droplet fluid as they mix allows for the number of cells in each droplet to be manipulated, to the point where just one is present. Not only does this technique have a low sample requirement, but it has high sensitivity and multiplexability, and can be done at high speed.
Droplets can be split or merged to facilitate downstream analyses. They can also be monitored over time, revealing the dynamics of their contents. For example, a pair of cells encapsulated together, such as a virus and its host cell, could provide insights into cell-cell interactions. The droplets can be lysed to extract their contents, while differences in metabolite concentrations between droplets can indicate the phenotypic variability between cells in a sample.
Then there’s single-cell sequencing, which investigates genetic and transcriptional heterogeneity of cells in a population. In single-cell RNA sequencing (scRNA-seq), cells are isolated into droplets, tagged with unique barcodes and their pooled RNA is transcribed and sequenced. This quickly tells us which genes are being expressed in each cell.
Early on, scRNA-seq was limited to cell populations with large numbers. “But then if you have to study rare populations, this is problematic because most droplets are empty,” says Miguel Xavier, a researcher at the International Iberian Nanotechnology Laboratory. Advances in scRNA-seq now make it possible to sort only occupied droplets, allowing researchers to study individual cells within rare cell populations. “Even if you’re looking at a population of 100 cells, like circulating tumour cells, you would be able to sequence all of them,” says Xavier.
This unlocks a new utility of droplet microfluidics in drug development. A major hurdle to developing therapies targeting circulating tumour cells is how rare yet diverse they are. If researchers can trap these cells in droplets, “then you can test different drugs or different concentrations of drugs for these cell populations”, says Xavier.
Other advances could further improve the performance of microfluidic single-cell analysis. In droplet microfluidics, detection relies on highly specific fluorescent labels, which limits what metabolites can be analysed. And in the future, AI-based imaging methods could enable label-free detection of droplets, automating ultrafast detection of all kinds of metabolites. Perturb-seq is a technique that combines CRISPR gene editing with single-cell RNA sequencing. “With Perturbseq,” says Di Carlo, “you can link functional changes in libraries of cells with downstream transcriptomic changes.” It illustrates the cellular responses to complex gene modifications and the regulatory pathways involved.
The cellular context
What cells do depends on their environment, including the extracellular matrix and their location in the tissue. For instance, cues from the extracellular matrix determine how and when a stem cell differentiates. While microfluidic technologies can isolate single cells, they usually don’t account for signals from the extracellular matrix.
As a result, existing microfluidic single-cell analysis leaves out many crucial details. “Except maybe some circulating immune cells, single cells are not in isolation in our body. Most cells are close to proteins and extracellular matrix and other cells,” says Di Carlo. New developments like producing single-cell microgels with droplet microfluidics are addressing this limitation.
For example, researchers are placing single cells inside little pockets in hydrogel, a biocompatible, cross-linked polymer. “Hydrogels mimic the native tissue environment,” says Castro Johnbosco, a PhD researcher at the University of Twente. In a 2024 study published in the journal Advanced Functional Materials, Johnbosco and his colleagues provided biomechanical cues to stem cells encapsulated in single-cell hydrogels. “You give each cell a unique microenvironment and by tuning it, you can regulate the stem cell phase,” he explains.
The location of different cells along a tissue also contributes to cellular heterogeneity. This is why there is a push towards spatial omics, an approach to combine genome or transcriptome analysis with mapping the position of individual cells. “Being able to look at the transcriptomics of single cells and combining them with the spatial information within the tissue is unbelievably powerful,” says Xavier. He adds that for cancer, it could expand our knowledge on how the disease develops and forms, the tumour microenvironment and metastasis.
Researchers are integrating spatial indexing methods with droplet microfluidics to advance highthroughput spatial transcriptomics. This has the potential to uncover disease and developmental trajectories. Meanwhile, there are efforts to replicate this approach for other omics technologies, such as lipidomics or proteomics. Microfluidic channels and droplets can also be made small enough to trap subcellular components. “Now the field is looking into single exosomes and single extracellular vesicles,” says Xavier. As with single-cell analysis, this research aims to analyse small populations of these exosomes and extracellular vesicles, maybe even one at a time.
From lab-on-a-chip to particle-on-a-chip systems
Lab-on-a-chip systems are one of the most translated formats of microfluidics for the analysis of biological samples. Here, as the term suggests, the cells are isolated and analysed in the same device. Cells from a tissue are typically cultured together in what are known as organ-on-a-chip devices. These are highly multiplexable, allowing for rapid screening of a large number of drug candidates. Now, lab-on-a-chip systems are being adapted for single-cell omics analyses. This opens up a host of clinically relevant use cases. For example, hundreds or thousands of antibodies could be simultaneously screened and their impact on cellular processes assessed. Yet some scientists are turning their attention to even finer details.
With the lab-on-a-particle approach, “The idea is using a particle, usually a hydrogel, that allows you to put cells or molecules onto it and compartmentalise a reaction so you can measure things very precisely,” explains Di Carlo. Another benefit is that reactions can be performed on the particle surface itself, unlike other microfluidics approaches where another component with a surface has to be added to the system.
Moreover, with lab-on-a-chip devices, as well as microfluidics more broadly, there is little room to innovate on top of an existing platform: each microfluidic single-cell analysis method has its specific assays and instrumentation. However, Di Carlo says, “In research, everyone is doing something slightly different and they want to do their assay a little bit differently.” Lab-on-a-particle approaches offer this flexibility – and they can usually be done using standard instrumentation found in life science research labs, Di Carlo adds.
Single-particle analysis can be achieved with droplet microfluidics or other approaches, as well. However, you would need to shrink the microfluidic channel size from, say, 20-by-20 micrometres to 2-by-2 micrometres. “The technology is there, but it needs to be adapted to the particles you are analysing,” says Xavier.
Manufacturing therapies
Microfluidic single-cell analysis could translate into applications spanning oncology, immunology, regenerative medicine and more. It will yield highthroughput insights into disease mechanisms and new biomarkers, thereby accelerating therapy development. However, if they are to be a commercial success, they’ll need to be cheaper to make. Presently, microfluidic chips are made with expensive materials in precision manufacturing facilities. Researchers are investigating designs that use inexpensive materials and more standard manufacturing processes, like 3D printing.
But the chip, or whichever microfluidic design traps the cells or particles, is only part of the picture. “People refer to a lab-on-a-chip but it’s more often a chip in the lab,” says Xavier. A range of specialised equipment, such as precision pumps and imaging systems, is required for microfluidic analysis. This makes operating microfluidic devices quite complicated and not very translatable for many labs. “What is missing, and people are working in that direction now, is standardising these devices so that they are compatible with common lab equipment,” says Xavier.
Finally, microfluidics could address the growing need to manufacture custom cells. Cell therapies, for example, can treat a range of diseases but manufacturing or performing quality control on them is often exorbitantly expensive. Microfluidics could help scale these efforts. “There could be a nice merger between new medical devices and new microfluidic approaches to perform these analyses,” Di Carlo says. “And lower the costs for cell therapies.”
