Jeffrey Perkel has been a scientific writer and editor since 2000. He holds a PhD in Cell and Molecular Biology from the University of Pennsylvania, and did postdoctoral work at the University of Pennsylvania and at Harvard Medical School.
When researchers want to know what cells are doing, they look to their transcriptional output. In the old days—that is, a decade ago or so—DNA microarrays and qPCR were the go-to tools for such studies. Today, RNA-Seq is dominant.
These methods are population-based—they reveal the transcriptional profile of the average cell in a sample. Increasingly, though, researchers are recognizing that such strategies are less than ideal, at least in some situations.
“Individual cells within the same population may differ dramatically, and these differences can have important consequences for the health and function of the entire population. Experimental approaches that only examine population-level characteristics can obscure these crucial differences,” explains the website for the Single Cell Analysis Program, a $90 million, five-year effort by the U.S. National Institutes of Health’s Common Fund “to accelerate the development and application of single cell analysis across a variety of fields.”
Researchers have over the past few years developed tools to downsize several 'omics disciplines to the single-cell level, including transcriptomics. Some methods are based on next-generation DNA sequencing, others rely on microscopy or flow cytometry. But all of them are exposing the heterogeneity of seemingly uniform cell populations as never before.
RNA FISH
Musa Mhlanga, principal investigator at CSIR Biosciences in Pretoria, South Africa, and at the Institute for Molecular Medicine in Lisbon, Portugal, studies the intricacies of gene regulation and its relation to genomic structure and transcription. In effect, he is interested in how the physical arrangement of genomic DNA in the nucleus impacts its expression.
In one recent study, Mhlanga's team investigated the impact of physical contact on the co-expression of three NF-κB-regulated genes distributed across two chromosomes. Previous population-level analyses using chromosome conformation and capture (3C) techniques combined with next-generation DNA sequencing had demonstrated that the three loci form a “multigene complex” following tumor necrosis factor-alpha stimulation, suggesting that their interaction (through chromosomal looping) was required to turn the genes on, Mhlanga says.
To test that hypothesis directly, the team used TALEN-based genome editing to disrupt those interactions in human umbilical vein endothelial cells, then examined cell by cell whether the regions were colocalized and the genes transcribed. Their method for making those assessments was RNA fluorescence in situ hybridization (FISH) [1].
In RNA FISH, each copy of an RNA molecule in a cell is decorated with multiple fluorescent tags. Under a microscope, transcripts appear as discrete dots, and by counting them, researchers can quantify precisely how many RNA molecules are present.
At least two labeling strategies exist. Biosearch Technologies’ Stellaris® RNA FISH strategy —which Mhlanga used in his study—hybridizes 40 or more short oligonucleotides along the length of each transcript, each labeled with a single fluorescent dye, thus amplifying the signal from each molecule some 40 times. Affymetrix’s ViewRNA™ strategy is based on branched DNA (bDNA) signal amplification. Here, the RNA-binding probe is not labeled; instead, it is connected by a bridge to a scaffold on which a series of amplifier sequences and up to 400 labels can be hybridized. “The analogy is of building a tree, starting with the trunk, then putting on branches, and you decorate that like a Christmas tree with labels,” explains Corina Nikoloff, senior product manager for ViewRNA assays at Affymetrix.
RNA FISH is a low-throughput approach. Researchers typically look at a relatively small number of cells and at a handful of transcripts per cell. In part, that’s because RNA FISH requires high magnification, typically 60x or more, and that in turn means small fields of view and often, samples under oil. The problem, explains Marc Beal, director of corporate development at Biosearch Technologies, is that it’s difficult to know if the few cells being studied are representative of the population, or the exception.
To address that problem, Biosearch is field testing a new imager called StellarVision®. StellarVision, Beal says, uses a strategy called “synthetic aperture optics”—exploiting light interference patterns and computer algorithms rather than microscope objectives—to image hundreds of cells simultaneously with a 20x field of view but 100x magnification, and all without oil. “For the first time, you’re able to see all the cells, not just one or two,” Beal explains. “So now you don’t have to worry about what’s going on north, south, east or west [of the cell you’re looking at], and you’re not under oil.”
According to Beal, the company has had beta units at the University of Pennsylvania and at Baylor College of Medicine since November 2014. The system should be fully available some time this quarter, he says, with an anticipated price tag of about $150,000 to $200,000.
Affymetrix also has developed a way to make RNA FISH higher throughput. Its PrimeFlow™ assay extends its ViewRNA approach to flow cytometry, albeit without the absolute quantitation. The method, says Dylan Malayter, senior product manager for PrimeFlow at Affymetrix, enables researchers to simultaneously measure both protein and mRNA content in single cells. For instance, suppose a researcher wanted to identify a particular cell subpopulation and quantify a transcript only in those cells. “You can do that with simple gating.”
As for Mhlanga, his results demonstrated that chromosome contact is indeed required for transcriptional activity. In the case of the NF-κB-regulated genes, chromosomal contact occurs in only about 5% of the cells he examined, and it is in those cells that the transcriptional coregulation was occurring. “We would never have been able to see that and to show that mechanism if we hadn’t done those experiments at the single-cell level,” he says.
Single-cell RNA-Seq
An alternative strategy for single-cell transcriptomics is RNA-Seq. Unlike RNA FISH, which measures just a few transcripts, RNA-Seq surveys all or most of the transcriptome at once. Also, where FISH requires researchers to know in advance which transcripts to study, RNA-Seq is unbiased. On the other hand, RNA FISH reveals a transcript’s subcellular location, but RNA-Seq does not. “Single-cell RNA-Seq is a wonderful complement to FISH,” says Candia Brown, director of the single-cell genomics business at Fluidigm.
The challenge with single-cell RNA-Seq is the single-cell part, specifically, isolating individual cells and preparing sequencing libraries. That’s where Fluidigm’s C1™ System comes in. The C1 is a microfluidics-driven system that enables researchers to capture, lyse and prepare libraries from up to 96 live cells, in parallel, in nanoliter reactions; a new in-development design will be capable of greater than 750 cells, Brown says. (Fluidigm has authored an ebook on single-cell preparation, available here. (PDF))
Aviv Regev of the Broad Institute of MIT and Harvard, in collaboration with researchers at Fluidigm, has used the C1 to prepare sequencing libraries from 1,775 dendritic cells as the immune system ramps up. The analysis identified a minor cell subset (just 2 of 1,775 cells) that displays a “precocious” antiviral gene-expression profile well in advance of other cells [2].
Regev’s study generated 4.5 million reads per cell on average. But a separate study coauthored by researchers in Arnold Kriegstein’s lab at the University of California, San Francisco, and at Fluidigm, suggests as few as 50,000 reads per cell may be sufficient, Brown says, at least if the goal is simply differentiating cellular subsets [3]. “It depends on the biological question,” she explains. “If you’re looking at tumors, they may need to sequence a little bit deeper”— perhaps up to 1 to 5 million reads per cell.
For the researchers who may want to give single-cell analysis a try, Brown offers this advice: The field may seem intimidating, but researchers needn’t be afraid. For one thing, she says, grant agencies seem to like the “novel point of view” that single-cell analyses bring. But also, as technology advances, barriers to entry fall. “The sophistication of the tools we and others are developing are making this an area that is intellectually a frontier, but with tools to make it accessible,” she says.
References
[1] Fanucchi, S, et al., “Chromosomal contact permits transcription between coregulated genes,” Cell, 155:606-20, 2013. [PubMed ID: 24243018]
[2] Shalek, AK, et al., “Single-cell RNA-seq reveals dynamic paracrine control of cellular variation,” Nature, 510:363-9, 2014. [PubMed ID: 24919153]
[3] Pollen, AA, et al., “Low-coverage single-cell mRNA sequencing reveals cellular heterogeneity and activated signaling pathways in developing cerebral cortex,” Nat Biotechnol, 32:1053–8, 2014. [PubMed ID: 25086649]
Image: CDKN1A exonic RNA, detected using Stellaris RNA FISH probes. Courtesy of Biosearch Technologies.