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 which proteins are present where in a tissue, they typically turn to antibodies. But that won’t work for unbiased molecular surveys, as researchers must decide in advance which antibodies to use. Plus, it’s completely impractical to use antibodies to scan for hundreds if not thousands of possible molecules, if for no other reason than cost.
One emerging solution to this problem is imaging mass spectrometry (IMS)*. IMS scans a tissue section point by point, amassing a dataset that correlates molecular species with their spatial positions. The result is like those transparent informational overlays sometimes found in geographic atlases that trace population, industry, demographics and so on, only in this case, the images specify molecular abundance (see, for instance, this video).
In 2007, Richard Caprioli, the Stanford Moore Chair in Biochemistry at Vanderbilt University School of Medicine, likened IMS images to digital photography, except that instead of the usual three color channels IMS yields thousands—one for each spectral peak. That definition still holds, but the images today are better than ever. What’s changed, in effect, is the maximum possible pixel resolution of the images researchers can take, and the variety of spectral “cameras” that can be used to generate them.
Differentiating features
IMS supports three basic applications, says Jim Langridge, health science director at Waters: Basic biochemical research, drug discovery (studying drug and metabolite distribution in tissue) and histopathology.
But the technology is rapidly evolving, and multiple approaches now exist, each with their own plusses and minuses. Some IMS approaches, for instance, require sample preparation prior to analysis, others don’t. Some favor large molecular weight species like proteins, while others excel at smaller compounds. And some operate in the open air (ambient conditions), while others require a vacuum.
That latter distinction isn’t just a matter of convenience, says Greg Kilby, director of molecular imaging and bioanalytical services at Protea Biosciences, which commercializes LAESI technology and offers it as a service. Some of his customers, for instance, study molecular distribution in contact lenses. These lenses are incompatible with vacuums, as the lenses get brittle and fall apart. “We’ve had a number of customers come to us saying they are looking for a solution, and MALDI [which often requires a vacuum] explicitly cannot be used,” he says.
R. Graham Cooks, the Henry Bohn Hass Professor of Chemistry at Purdue University, suggests breaking down IMS methods by how they ionize and vaporize a sample. SIMS and its kin use an ion gun, liberating molecules by the brute force of collision with such particles as ionized C60, gold and bismuth-3, while MALDI, LAESI and NIMS rely on laser energy. DESI and LESA are solvent-based, using liquid solvents to extract molecules from the sample surface. (A fourth imaging strategy, rarely used in biological research, involves plasma ionization.)
Key IMS methods
SIMS
SIMS provides phenomenal spatial resolution—on the order of 100-nm or so—and the ability to map molecular distribution in three dimensions, says Nicholas Winograd, the Evan Pugh Professor of Chemistry at Pennsylvania State University. But the technique is too energetic for many biomolecules, which shatter upon ionization, so it’s used mostly for lipids and small metabolites. Also, the pixels in SIMS are so small that sensitivity can suffer for lack of material. (In general, sensitivity tracks spot size. Caprioli advises a balancing act: “Use the biggest ablation size that answers your biological question.”) And then there’s cost: SIMS is expensive compared to other mass spec approaches, Winograd says, with some instruments costing $1.5 million apiece. “For the biologist who wants to do routine spectra, SIMS is not the technique to use.”
MALDI
Probably the most widely used method, MALDI offers resolution on the order of five- to 10-micron with just a little tweaking to off-the-shelf systems, Caprioli says—by installing a lens and pinhole in the laser line, for instance—and submicron resolutions are possible. The technique also supports a wide range of molecular classes, from proteins (Caprioli’s forte) to small molecules.
MALDI requires a matrix to be deposited on the sample prior to analysis; the matrix absorbs laser energy, but it also extracts compounds from the surface and co-crystallizes them with the matrix itself. That deposition step adds time to experiments and can be challenging to optimize and reproduce. (MALDI matrix-deposition robots, such as Bruker Daltonics’ ImagePrep and Sunchrom’s SunCollect, automate this process.) Also, MALDI usually (but not always) takes place in a vacuum, and matrix ions can overshadow the low molecular weight end of the spectra, whcih can complicate small-molecule analyses.
Still, many do use MALDI IMS for that application. At ImaBiotech, a service provider using MALDI-coupled instruments from Bruker Daltonics, researchers use quantitative high-resolution IMS to focus on xenobiotic, small-molecule and peptide distribution and toxicity profiles for drug discovery, says CEO Jonathan Stauber. “We don’t deliver services for larger molecules at this stage.”
LAESI
LAESI is an ambient, laser-based method that requires no matrix. Instead, the technique uses a mid-infrared laser tuned to the absorption properties of water, so no sample prep is required—though a hydrated sample obviously is. Another pro: LAESI ablates tissue as it scans, says the technique’s inventor, Akos Vertes of George Washington University, meaning it (like SIMS) can be used for molecular depth profiling. And the technique’s spatial resolution is high enough that Vertes uses the technique to image not evenly spaced points on a grid but individual cells themselves—an approach he has used to trace pigment distribution in onion tissue. “I believe this is a much more biologically relevant way of imaging,” he says.
NIMS
Another laser-based technique, NIMS uses a nanostructured porous silicon surface to volatilize samples. The tissue section sits on a nonstick fluorinated polymer, like Teflon, atop the structured silicon surface. That nanostructured surface obviates the need for a matrix, because its structure prevents laser-induced heat from dissipating. Instead, energy is effectively trapped in tiny wells, exciting the sample enough to ionize and volatilize it. The key to NIMS, though, is the fluorinated polymer, says Gary Siuzdak, director of the Center for Metabolomics and Mass Spectrometry at the Scripps Research Institute, whose lab invented the approach; the fluorinated polymer’s nonstick properties enable the technique to achieve ionization with relatively little input energy and thus relatively little fragmentation. “Just as an egg easily slides off [a Teflon-coated] pan, the molecules are easily released from the fluorocarbon surface using very little laser energy,” Siuzdak says—about three-times less energy than MALDI. NIMS targets are not commercially available, but instructions to prepare them are available on Siuzdak's web site.
LESA & DESI
LESA and DESI are both ambient solvent-based ionization methods. In LESA, commercialized by Advion, a droplet of solvent is placed on the sample at a specific position, extracting compound material, then taken up and injected into the mass spectrometer. DESI (available from Prosolia) sprays the liquid on the sample, where it pools and extracts material beneath the puddle. Subsequent droplets cause that liquid extract to splash up into the mass spectrometer for analysis. DESI is best for relatively small molecular weight materials, says Cooks, such as lipids and drug metabolites, and offers spatial resolution on the order of 25 μm.
IMS meets histopathology
Recently, groups led by Garry Nolan at Stanford University and Bernd Bodenmiller at the University of Zürich independently reported new methods that effectively represent a cross between IMS and histopathology [1,2]. Instead of unbiased imaging approaches, both teams used panels of lanthanide-tagged antibodies to probe specific proteins. Up to 100 antigens presumably could be screened simultaneously via this approach, though Nolan’s team tested only 10 and Bodenmiller’s, 32. Bodenmiller detected those metals in a modified CyTOF mass cytometer from DVS Sciences, while Nolan’s team used a high-priced, high-resolution SIMS-coupled magnetic sector mass spec to achieve the same result.
Michael Angelo, the post-doc who led the Nolan lab team, says their study, which profiled antigens in formalin-fixed, paraffin-embedded (FFPE) breast-cancer tissue with 200- to 300-nm resolution, was essentially a proof of principle, demonstrating the technique (called MIBI) could be applied to the millions of archived FFPE samples pathologists have at their disposal. That’s assuming MIBI can be adapted to more user-friendly, less-expensive hardware. “That’s the big take-home point,” he says. “The conceptual ability is right there. The only reason we can’t deploy this right now is an instrumentation issue.”
Key considerations
When selecting an IMS strategy, there are several variables to consider. Perhaps the most significant is the nature of the analyte, says Vertes. Proteins are large molecules, and they’re likely to stay in place during sample processing, so they are amenable to MALDI. But small-molecule signals may blur during sample prep, because the molecules themselves are so mobile and soluble in solvent. Ambient ionization approaches, such as DESI or LAESI, may work better in such cases.
What mass range do you anticipate probing, and at what resolution? IMS offers no fractionation, so it’s more akin to sampling crude lysates than LC-MS, says Kilby. (The exception is inline ion mobility separation, which resolves molecules by shape, for instance in Waters’ Synapt® MALDI mass specs.) Will simple spectra suffice, or do you anticipate needing tandem MS capabilities for molecular identification? Do you require high mass accuracy and mass resolution, or will lower-performance equipment suffice?
Larger molecules require a wider mass range, which is why time-of-flight analyzers are so frequently used for protein work. As for pixel resolution, says Caprioli, as with computer monitors, “the lower the number of pixels, the cheaper it is. There’s a trade-off there. [But] most of us like as many pixels as we can get.” On the other hand, the smaller the pixels, the less material released per point (and the longer it takes to complete an experiment).
Finally, do you need 3D capabilities? If so, would you prefer to image slices or let the instrument handle that for you?
The answers to those questions obviously depend on the question being asked. And IMS methods can be mixed-and-matched to maximize data. (One recent study combined MALDI and LESA IMS on a Bruker Daltonics 12 T solariX FT-ICR mass analyzer to investigate drug distribution in rat lung, for instance [3].)
But ultimately, says Sören-Oliver Deininger, MALDI imaging market manager at Bruker Daltonics, it’s not the image itself that’s the goal in IMS. The payoff is in the comparisons such data enable. “A real study doesn’t stop at the image itself,” he says. “The image is just a tool to find the right molecule in the long run.”
References
[1] Angelo, M., et al., “Multiplexed ion beam imaging of human breast tumors,” Nat Med, 20:436–42, 2014. [PubMed ID: 24584119]
[2] Giesen, C., et al., “Highly multiplexed imaging of tumor tissues with subcellular resolution by mass cytometry,” Nat Methods, 11:417–22, 2014. [PubMed ID: 24584193]
[3] Tomlinson, L., et al., “Using a single, high mass resolution mass spectrometry platform to investigate ion suppression effects observed during tissue imaging,” Rapid Commun Mass Spectrom, 28:995–1003, 2014. [PubMed ID: 24677520]
(*) Abbreviations used in this article: DESI, desorption electrospray ionization; FT-ICR, Fourier transform ion cyclotron resonance; IMS, imaging mass spectrometry; LAESI, laser ablation electrospray ionization; LC-MS, liquid chromatography-coupled mass spectrometry; LESA, liquid extraction surface analysis; MALDI, matrix-assisted laser desorption ionization; MIBI, multiplexed ion beam imaging; NIMS, nanostructure imaging mass spectrometry; SIMS, secondary ion mass spectrometry
Image: A 3D depth profile created using LAESI. Courtesy of Akos Vertes.