Demanding more from flow cytometry means putting more into the design of an experiment. From controlling for spectral overlap and orders-of-magnitude differences in fluorescence among large numbers of colors, to making sure only single living cells are counted, there’s a lot to think about—even if someone else ultimately operates the machine. Here are some of those considerations.
The match game
When designing a flow experiment, counsels Aja Rieger, the University of Alberta’s flow cytometry core manager, the first thing is to know your machine, your optical setup, and the fluorochromes that are available on your machine. Next is to know the biology of your system—what kind of cells you’re working with, what you want to detect on those cells, approximately how much of whatever you’re looking at those cells express, if it’s intracellular, extracellular, nuclear, and so forth. Then “use that knowledge to start pairing your markers of interest with a fluor.”
To do the latter, Kelly Lundsten, business segment manager for advanced cytometry at BioLegend, first breaks the markers down into tiers. “Primary tiers are basic phenotypic markers—things that don’t really change,” like CD4 and CD8 for categorizing T cells, for example. These are likely to yield a yes/no answer.
In the secondary tier are things like activation and exhaustion markers, or “second-level phenotypic markers that get you a higher-level phenotype, like T helper subsets.”
And “the tertiary tier would be your more interesting question—the thing you’re probably doing the entire assay for … and that is going to vary the most between your patient samples,” she continues. Things like transcription factors or cytokine expression, for which quantification is likely to be important.
As a broad generalization, tertiary markers— “the most important things in your panel” —tend to have the lowest expression, says Rieger. “You’re going to want to give these the brightest possible fluorochrome, because they are generally the hardest to detect for most panels.” Conversely, highly expressed or qualitatively invariant markers can often be paired with dimmer fluors.
But generalizations have exceptions, points out Sukhwinder Singh, technical director of the Rutgers New Jersey Medical School flow cytometry and immunology core laboratory. He cautions against using dyes such as phycoerythrin (PE) in some instances “because you’re going to see a lot of spread and such,” requiring significant compensation. Along similar lines, Singh says to pick colors that are far apart when possible, and have different fluors excited by the different lasers.
Panel construction tools found on many provider’s websites are good places to start. They consider their available antibody-fluor conjugates as well as the instrument’s excitation laser lines and filter sets, and suggest the optimal combinations for a given set of antigenic markers.
“The last step … is to test to make sure everything seems to be working according to the biology you expect,” implores Rieger, because “there will always be little things that will affect that panel and might require you to switch out markers.”
All about control
To assure that the biology is working according to expectations, and to get the data you’re looking for, it’s important to run controls (and in flow cytometry that goes beyond just biological controls).
Unstained cells alone establish the baseline of autofluorescence. “There are a lot of mechanisms in cells that have some inherent autofluorescence,” points out Joel Sederstrom, director of flow cytometry at Baylor College of Medicine. Single positive controls determine how much each fluor spills into other channels of interest—these “are controls for every channel but what it’s positive for.” Meanwhile isotype controls “are no longer considered a real control. They really only tell you how good your blocking is.”
As with any bioassay, it's also important to monitor for the experimental conditions themselves. Is there a difference between treated and non-treated cells?, for example, and How do timing and dosage affect the result?
Assay controls deal with background and biological considerations, but there are other types of controls crucial to running a meaningful flow cytometry experiment. Calibration controls assure the instrument is performing consistently. Compensation controls deal with spectral overlap among different fluorophores. And analytical controls help establish gating parameters, says Lundsten.
The Bottom Line
1. Know your instrumentation and the reagents available for the optical setup.
2. Know the biology of your system.
3. Pair dim, variable, and important antigens with bright fluors. Pick colors that are far apart, and make use of all the available laser lines. Use a panel builder to help.
4. Use a viability dye to screen out dead cells.
5. To assure a single cell suspension, try to avoid Ca++ and Mg++, and use EDTA, in the medium.
6. Controls, controls, controls. But don’t rely on isotype controls for gating.
“That’s kind of the step-wise fashion of applying controls: set up your instrument; know how much your fluorochromes overlap and be able to correct for that; be able to have an intelligent look at your data by understanding the biological considerations and those controls,” she continues. “And then probably the most neglected are the controls that will help you to do your analysis correctly, and those are fluorescence-minus-one [FMO] controls.”
As for the compensation controls, they should be exactly matched to the fluors being sampled. Tandem fluors, like PE-Cy5, should additionally be compensated for by the same antibody. And fluors should always be at least as bright as samples. “They’ll have compensation problems because their compensation control is dimmer than their actual sample,” says Rieger. “The same thing with viability dyes: People will bring a live cell to compensate for their dead cell assays, but there are no dead cells so you can’t compensate anything. You always want to think through those controls and make sure that they’re going to represent what’s actually happening in the experiment, in terms of expression of your markers.”
Prepping for flow
Before you even get to the instrument, it’s important that samples are ready to be read. You should of course make sure the cells are properly cultured, assays are designed and scheduled to reflect the biological time course, and that reagents are titrated, points out Singh.
“You want single, intact cells in the end that you can analyze,” emphasizes Sederstrom. So using things like live/dead viability indicators is critical. You usually want to make sure the medium doesn’t contain calcium or magnesium, and maybe use EDTA, because “calcium is necessary for integrins to bind, and integrins mean the cells can stick to each other.” Protein in the medium helps keep the cells in suspension. He also recommends at least pulsing with DNAse “because when cells die necrotically they eject their DNA,” causing cells to stick together.
New flow cytometry instrumentation, reagents, and software promise simultaneous querying of ever-increasing numbers of parameters. With increased multiplexing comes increased complexity, and so to take maximum advantage it behooves researchers to give ever more consideration to sample prep and panel design. All while never forgetting that it’s the controls that give credence to the data.