Imaging flow cytometry (IFC) brings together some of the best aspects of flow cytometry and of automated digital fluorescence microscopy. Rapid interrogation of many thousands of cells allows for robust statistics, so discrimination of rare events is easily carried out with this methodology. With microscopic imaging of cells, differently hued arrays of differentially bright pixels afford discrimination of morphological features, and fluorescent intensity (FI) and localization, at a subcellular level of detail. IFC enables researchers to perform detailed characterizations of subpopulations of cells that yield far more information than could be had from using the two techniques separately.
Here we offer a sampling of some of the applications—for example, examining intracellular trafficking, the formation of immune synapses and cell signaling—that IFC is being used for today.
How it works
There are commercial IFC instruments, such as FlowCam, that can analyze small particles and organisms, but they lack the resolution and number of parameters for the type of biomedical analyses that the MilliporeSigma Amnis instruments are capable of. Henceforth, “IFC” implies use of Amnis instruments (or home-brew instrumentation with similar capabilities).
In IFC, particles suspended in sheath fluid are interrogated as they stream through a flow cell at a rate of up to 5,000 per second.
Light exits the flow cell through (up to) a 60x objective and enters a multispectral decomposition unit (like a prism) that separates it into different colors. Images—up to 12 per cell—are recorded as individual pixels, one row at a time, by a CCD camera (or two), tracking the cell as it traverses the flow cell. “It’s very similar to your standard flow analysis. You create dot plots or histograms, and you gate on those,” explains Amnis product manager Sherree Friend. “It’s a little easier, because you can actually click on a dot and see the image and decide where your gating should be.”
Gating—determining the boundaries of bins used to classify the cells—can be accomplished on the basis not just of overall FI and light scattering, as it is in traditional flow cytometry, but also based on a cell’s aspect ratio, diameter and a host of other morphological features.
About half of the University of North Carolina Flow Cytometry Core Facility IFC users “want to do the same thing they do on a flow cytometer. They just want to see images, to get a visual,” says technology manager Sébastien Coquery. They want, for example, to make sure they are gating properly on large singlets vs. doublets. “Instead of guessing where one starts and the other stops, you can just click on individual dots and see,” Coquery explains. The users also want a better understanding of where the staining lies, and how it might change upon treatment—is it uniform? Is it on the cell surface? Is it in the nucleus? Is it in the Golgi or endoplasmic reticulum (ER)? “And they can merge [the channels] together, like you would do in microscopy,” he says.
The other half of Coquery’s users want to know more; they want robust statistics. For example, he says, “if they treat the cell with a compound, are there any changes in the localization of the staining or a change in the shape of the cell?” In addition, “there are a lot of people doing phagocytosis, putting bacteria in with neutrophils or monocytes and quantifying the uptake with the software.”
Putting the statistical power of IFC into perspective, for a particularly expressed cell within a population occurring at a frequency of 0.1%, it would take an average of 10 microscope slides with 100 cells on each to find a single cell.
Software
The Amnis instruments’ powerful IDEAS software contains “wizards that can step you through an internalization assay, or a nuclear localization assay, or an apoptosis assay, or a shape-change assay,” says Friend.
When there is no validated assay, such as for looking at cell-cell interactions, the software’s new Feature Finder wizard can discriminate between populations. “You need to identify that location [of the synapse], and you might need to calculate a new feature that isn’t calculated by default … for example, actin accumulation … and pick ‘truth populations,’” Friend explains. “The Feature Finder will actually calculate additional features and determine which feature will separate those two populations the best. … And then you validate that based on your whole population at the end.”
It is “almost a blind search,” remarks Natasha Barteneva, director of the flow and imaging cytometry resource in the Program in Cellular and Molecular Medicine at Boston Children’s Hospital and faculty associate at Harvard Medical School.
“You can create new features, and you can create new masks. You can combine masks. You can combine features. You can divide, get ratios, get sums, basically create a new equation,” Barteneva says. It is simple to use, with no scripting involved. Other software, such as ImageJ and MetaMorph, also allows users to “do a lot of things,” she says, but it requires some understanding of programming. “You’re writing scripts, or you’re borrowing a script that someone already wrote.”
The introductory chapter to Barteneva and her co-editor’s 2016 book on IFC calls Feature Finder “one of the most important and user-friendly recent developments in imaging analysis software” [1].
Downsides?
For example, “in terms of the resolution of the images, it will never be as good as a stand-alone microscope,” Coquery points out. Jessica Back, associate director of the Wayne State University Microscopy, Imaging, and Cytometry Resources Core, notes that “if you want to look at real morphology like you see in a 3D culture, or something that looks like a z-stack on a microscope, you’re not going to get that with this system. It’s not intended for that.”
As with many hybrid systems, IFC cannot always do everything (or do everything as well as) the instruments that it integrates.
Along the same lines, you can see the shape of cells such as neurons in a microscope. “They’re long and skinny,” Back says. “But when you put them into a flow-based system, they automatically round up because of surface tension. … You lose that kind of spatial arrangement.”
The Amnis instruments also lack the throughput of a high-end flow cytometer, which can acquire millions of events in under a minute. And it cannot physically sort or otherwise recover cells that have been analyzed.
Researchers may sort cells to enrich for a rare event and then look in a microscope to confirm the morphology, for example, or to confirm that some activity is happening in the cell. “You wouldn’t need to do that with an [IFC],” says Friend. “You can gate down on what cells you’re interested in. So if your goal is to image the cells, you can eliminate that whole step.” Of course, there is no substitute for sorting if the goal is to culture the subpopulation or to perform downstream biological analysis on it.
With many more than 700 peer-reviewed publications (including three books or dedicated journal issues) and users’ groups around the globe, researchers have certainly found uses for IFC. In fact, Barteneva predicts that “this is a technology which will change biomedicine in the next few years. It’s answering questions that cannot be answered by microscopy and cytometry.”
Reference
[1] Barteneva, NS, Vorobjev, IA (eds), “Imaging flow cytometry: methods and protocols,” Springer Protocols: Methods in Molecular Biology series, Humana Press, New York, 2016.
[NLM ID: 101676394]
Images: Shutterstock Images