Inclusion of companies in this article does not indicate endorsement by either AAAS or Science, nor is it meant to imply that their products or services are superior to those of other companies.
Akos Vertes, professor of chemistry at George Washington University in Washington, DC, uses mass spectrometry methods to quantify metabolites at the single-cell level. He offers a simple justification for the work: “Cell populations are heterogeneous. And these differences between individual cells in large populations sometimes become very important.”
Consider cancer stem cells, for instance. Or drug-resistant bacteria. Or the many subtly different neuronal cell types in the brain. It’s possible to detect those subpopulations in a bulk analysis, but it isn’t easy: Their signals tend to be swamped by the general population. And in any event, such analyses blur cell-to-cell distinctions, making it impossible to know which cells contribute what to the population. The only way to untangle that skein of competing signals is to make measurements cell by cell.
Other researchers turn to single-cell methods because they have no other choice. In microbiology, for instance, the vast majority of microbes cannot be cultivated in the lab, says Ramunas Stepanauskas, director of the Bigelow Laboratory Single Cell Genomics Center in East Boothbay, Maine. "We really don't know much about them, although they are the drivers of the global biogeochemical cycles; they inhabit every conceivable environment, including our own bodies, and they are a largely untapped source of novel natural products, bioenergy, and other applications." The best way to tap these organisms' capabilities, he says, is to sequence them one at a time, letting their genetic material tell researchers what Petri dishes cannot.
Fundamentally, the process of single-cell research isn’t difficult at all. Simply isolate a single cell, lyse it, collect its molecular innards, and analyze what you get. Of course, it’s not nearly so straightforward in practice.
“There are all kinds of pitfalls in single-cell analysis,” says Jim Eberwine, codirector of the Penn Program in Single Cell Biology at the University of Pennsylvania School of Medicine, who has taught a course on single-cell analytical techniques at the Cold Spring Harbor Laboratory since 2012 and has been studying single-cell gene expression in his own lab since the 1990s. According to Eberwine, whether a researcher is interested in a cell’s DNA, RNA, protein, or metabolites, the fundamental issue is always the same: “Can you detect your signal over the noise.” Researchers have devised a number of clever strategies to do just that.
Of single cells and amplification
The key to single-cell biology, of course, is isolating a single cell. This typically is accomplished using micromanipulation, microfluidics, or fluorescence-activated cell sorting (FACS), the method Stepanauskas favors.
One alternative, developed in the laboratory of Nancy Allbritton, chair of the joint biomedical engineering department at the University of North Carolina at Chapel Hill (UNC-CH) and North Carolina State University (NCSU), is the microraft array (MRA).
An MRA is an array of tiny transparent, magnetic, polystyrene elements, each small enough to fit into a tiny well on a plate or slide. Commercialized by Cell Microsystems, MRAs typically contain about 10,000 wells, Allbritton says, though some measure in the millions. Researchers plate their cells such that there are, on average, zero or one cells per element. They can then image the array immediately, or allow the cells to grow and develop complex phenotypes. Cells of interest are gently isolated using a microneedle to pierce the bottom of the array and dislodge the MRA element. Because they are magnetic, these elements are easily captured, at which point they can be analyzed or clonally expanded.
According to Allbritton, the system enables isolation strategies that might otherwise be impossible, such as isolating cytotoxic T lymphocytes based on their ability to kill target cells. In one study, UNC-CH collaborator Scott Magness, working with Allbritton, used an MRA to study intest-inal stem cells. Among other things, the team used the platform to investigate whether intestinal stem cells must physically contact other cells in order to reach their full potential. “What he showed quite clearly is they really need to be touching a Paneth cell to grow, prosper, and divide and get the highest outgrowth rates,” Allbritton says.
Vertes takes a lower-throughput approach to metabolomics. Using a sharpened capillary and a micromanipulator—a tool typically used in patch clamping and embryonic manipulation—his team aspirates a fraction of a picoliter of material from adherent cells stuck to the bottom of a culture dish and injects it directly into a mass spectrometer.
In that way, Vertes says, his team can quantify some 22 metabolites and 54 lipids from each of about 30 cultured hepatocytes—not the whole metabolome, of course, but enough to reveal for instance, the cells’ adenylate energy charge—a measure of cellular health. It was always possible to measure such variables at the population level, Vertes notes. But by going to the single-cell level, he can observe the distribution
of energy states, and correlate those values with such parameters as morphology or metabolic activity. “For every cell, we can tell [if it] was a healthy cell, a cell half-dead, or a cell programming itself to die,” he says.
Jonathan Sweedler, the James R. Eiszner Family Chair in Chemistry at the University of Illinois at Urbana-Champaign, has developed another novel approach to cellular isolation and analysis. His team disperses some 10,000 cells onto a microscope slide. They then stain the nuclei of those cells, localize them with a fluorescent microscope, and pass those coordinates to a laser for mass spectral analysis. Because the cells are widely dispersed, there’s never a question of whether a laser is hitting one cell or another, as can occur in traditional mass spec imaging. “We can actually do 100,000 cells this way in an afternoon if we wanted.”
Sweedler’s approach thus trades cellular context and interaction for throughput. He can quantify many more discrete cells than he otherwise might, but he loses any sense of where those cells were in relation to one another. That’s a common problem with single-cell approaches, Eberwine says, because it also potentially alters cell biology. “You’re removing the cells from their natural microenvironment, where they have the adjacent cell connections and the influence of the cellular environment.”
Indeed, using a method called “transcriptome in vivo analysis” (TIVA), Eberwine and his colleagues compared the gene-expression patterns in pyramidal neurons in intact mouse brain slices and dissociated cells. They found that neurons in intact tissues expressed 12,000 transcripts on average, compared to 16,000 when dissociated. That finding suggests that a cell’s capacity to express a gene is modulated by the physical signals it receives from its neighbors. The local microenvironment “act[s] as a brake for gene expression,” Eberwine says.
TIVA employs a caged, photoactivatable double-stranded oligonucleotide coupled to a cell-penetrating peptide and biotin moiety. When this “TIVA tag” crosses the cell membrane, the peptide is removed, locking the molecule within the cell. Laser irradiation then causes the two strands to separate (“uncaging”), revealing a sequence for capturing polyadenylated messenger RNAs (mRNAs) in the targeted cells. Finally, the biotinylated RNA-RNA hybrids are isolated on streptavidin beads, amplified, and analyzed via RNA sequencing (RNA-seq), allowing the team to quantify gene expression from locales ranging from specific cells to subcellular compartments. “We can activate it in the cell soma, we can activate it in the dendrites or in the nucleus of the cell, and we can do it in one or multiple cells,” Eberwine says.
At Harvard University, Xiaowei Zhuang, a Howard Hughes Medical Institute investigator and David B. Arnold Professor of Science, who invented the superresolution method of stochastic optical reconstruction microscopy (STORM), has also developed a technique for in situ single-cell transcriptome analysis. But unlike TIVA, her imaging-based method retains information on precisely where in the cell the transcripts were located. “We know that RNA are not uniformly distributed inside the cell,” Zhuang explains. “The local distributions of RNA are actually quite important for the establishment and maintenance of local cellular structures.”
Multiplexed error-robust fluorescence in situ hybridization (MERFISH) is an in situ hybridization-based approach for quantifying thousands of transcripts simultaneously. In one specific implementation, each transcript is assigned a unique binary code. The trick is to read that code bit by bit.
To do that, oligonucleotides complementary to each transcript are flanked by readout sequences corresponding to the different bits. This pool, which may comprise 100,000 oligos or more, is hybridized to the mRNAs in the fixed cells, “convert[ing] each cellular RNA into a combination of readout sequences,” Zhuang explains. Then, the cell is interrogated bit by bit using oligonucleotides complementary to the readout sequences. After each hybridization round, the cells are imaged, the fluorescent probes quenched, and the process repeats. Each fluorescent signal in any given round represents an RNA whose code reads “1” at that position in the corresponding binary code. By mapping these points over 16 rounds of hybridization and quenching, the system can quantify numerous RNAs in the cell, and map their subcellular location.
But according to Zhuang, multiplexing is only half the story. FISH has a low but measurable error rate, Zhuang says. Over 16 rounds of hybridization, that can add up. So, taking a cue from telecommunications, her team designed its codes such that each code differs from every other one by up to four bits. “That means you have to make four errors in one RNA in order to convert it to a different RNA,” she explains. Thus, even if a particular transcript is read out with a single error, the software can determine its correct identity, although RNAs containing two or more errors can be detected but no longer corrected, as it is no longer possible to unambiguously identify them. “We can actually do both error detection and error correction,” Zhuang says, “and that’s how we get very highly accurate measurements.”
Using this strategy, Zhuang’s team quantified as many as 1,000 transcripts in individual cells. But there’s no reason that number of transcripts cannot go higher, she adds. A 32-bit code, for instance, could easily encode 30,000 genes. Now she hopes to apply the technology to map the different cell types in the human brain and in cancer, as well as the subcellular distribution of RNAs within them.
Bernd Bodenmiller, assistant professor for quantitative biology at the University of Zürich, Switzerland, and Sean Bendall, assistant professor of pathology and immunology at the Stanford University School of Medicine, are both former postdocs of Garry Nolan, also at Stanford, who champions mass cytometry, a cross between flow cytometry and mass spectrometry that overcomes one of the key shortcomings of the former technique.
Traditional flow cytometry, constrained by overlapping fluorescent channels, is limited to about 18 simultaneous signals, or markers, though in practice, the real limit is even lower. Mass cytometry can theoretically resolve more than 100 channels, but is currently limited to about 44—the result of conjugation chemistry needing to catch up with the instrumentation itself.
Like flow cytometry, mass cytometry, commercialized by DVS Sciences as the CyTOF (recently acquired by Fluidigm), is a flow-based method that uses heavy metal-conjugated antibodies rather than fluorescent dyes to achieve its high information content. But in a pair of papers published in 2014, Bodenmiller and Bendall simultaneously demonstrated methods to turn a CyTOF analyzer into a mass spec imager, enabling a kind of high-throughput single-cell proteomics spatial analysis. “We built what you could call a mass cytometry-based microscope,” Bodenmiller explains, enabling “spatial systems biology or spatial proteomics.”
As with advanced mass spec imaging systems, this approach allows researchers to measure protein content cell by cell in situ. But, unlike traditional mass spec imaging, which typically detects the most abundant proteins, mass cytometry allows researchers to actually preselect the proteins they want to image.
Bodenmiller used his mass cytometry microscope to study the impact of tumor microenvironment on metastasis—for instance, the impact of macrophage infiltration. “I think here is the sweet spot of our imaging mass cytometry approach, because it’s antibody-based, so we can exploit prior knowledge to detect different immune cells or stromal cells, and we can use many antibodies to learn what these cells are actually doing.” Bendall and Nolan used their multiplexed ion beam imaging (MIBI) system to probe the histology of breast cancer sections.
Separately, Bendall and Nolan used traditional CyTOF mass cytometry to map the single-cell trajectories of B-cell development, using a set of 42-marker molecular profiles—everything from extracellular markers associated with B-cell development to intracellular signaling factors—and a novel algorithm called “Wanderlust” to trace how molecular signatures change as a lymphocytic stem cell differentiates into a mature B cell.
Though 40-plus parameters are impressive, researchers like Bendall and Bodenmiller hope to ultimately push that number even higher. All that’s required is efficient chemistry to conjugate the ions to the antibodies. Bendall jokes of a creeping condition among flow and mass cytometrists that he calls “PDS,” or “parameter dependency syndrome.” “At first you ask yourself, ‘Well, what the heck am I going to do with 40 parameters?’ And then after you’ve done this for a while, you’re like, ‘How am I [going to] fit everything I want to measure into 40 parameters?’” To put it another way, researchers never say no to more data.
Of course, those data describe just one particular aspect of single-cell biology. And therein lies one of the field’s primary challenges, says Eberwine. What’s needed now, he says, are methods for integrating these different datasets to gain a comprehensive, quantitative, and reproducible view of the cell. Work on that front is ongoing. “I think the information will be incredible,” he says.