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.
Eukaryotic tissue is, as we all know, not homogenous. It is built of different cell types, with distinct functions, morphologies and gene expression programs. Many of the most biologically interesting functions of the brain, for instance, are performed by neurons. But neurons come in several flavors—such as motor neurons and sensory neurons—and share cerebral real estate with other cells, like glia and astrocytes, as well.
Researchers are well aware of this fact. Yet all too often, they work at the level of tissues, measuring gene expression in the heart, say, or in a tumor, an approach that produces a molecular description of the tissue overall, not the cell type of interest.
“If you have a heterogenous tissue, you will get an average response of the mixed population,” says Chris Vega, marketing manager for life science research at Leica Microsystems.
In a 1996 Science paper, Lance Liotta, then at the National Cancer Institute (NCI), together with his colleagues described a solution to this problem, which they called laser capture microdissection (LCM). “The method,” they wrote, “entails placing a thin transparent film over a tissue section, visualizing the tissue microscopically and selectively adhering the cells of interest to the film with a fixed-position, short-duration, focused pulse from an infrared laser. The film with the procured tissue is then removed from the section and placed directly into DNA, RNA or enzyme buffer for processing.”
The technology Liotta and colleagues described was commercialized by a company called Arcturus, which is now part of Life Technologies. Three other commercial suppliers—Carl Zeiss Microscopy, Leica Microsystems and Molecular Machines & Industries (MMI)—have since developed related systems. For researchers who want to home in on specific cellular subsets, that’s good news. Says Liotta, “For the good of science, you need to microdissect. It doesn’t matter which system you use, it’s better to dissect and isolate homogeneous cell populations than to use heterogeneous populations.”
Commercial LCM systems
Liotta is co-director of the Center for Applied Proteomics and Molecular Medicine at George Mason University. The lab possesses three LCM systems, all from Arcturus: a first-generation, manual PixCell instrument; a newer Veritas system with a graphical-user interface for laser control; and the company’s newest system, a dual-laser ArcturusXT™.
The lab uses those instruments in its clinical trial work, Liotta says. In one example, Virginia Espina, director of the Center’s Clinical Proteomics Laboratory, microdissected tumor tissue from 100 breast cancer patients before and after anticancer drug treatment. She then extracted protein from those samples and applied them to protein microarrays to map the impact of the drugs on the amount and phosphorylation of nearly 60 protein biomarkers.
“We came up with a network map of how the tumors that responded or didn’t respond were different,” Liotta says. This is information he hopes may one day guide therapeutic decisions.
Others use LCM for molecular profiling of either RNA or DNA. “The primary downstream analysis platform with LCM customers is genomic analysis,” says Shirley Chu, Arcturus product manager at Life Technologies—mostly qPCR or microarray analysis of RNA, though next-generation genome sequencing and transcriptome analysis are also possible.
The ArcturusXT uses the same infrared laser capture microdissection technology originally developed by Liotta’s group at NCI. The IR laser gently bonds cells of interest to a photoreactive membrane sitting atop the tissue slice. Those cells then are collected by lifting the membrane, which is attached to a CapSure® LCM consumable—basically, the lid of an Eppendorf microcentrifuge tube.
That bond-and-lift approach works well for individual cells, collections of nearby cells or even subcellular fractions. But it’s less efficient for collecting large regions or for particularly challenging samples such as bone, says Chu. For those cases, the ArcturusXT also optionally includes an ultraviolet laser, which circumscribes a region of interest to excise it from the tissue section. (In this case, the IR laser is also used to bond the region to the membrane via discrete “glue points,” says Chu.)
Only Arcturus LCM instruments use an IR laser to “capture” cells; systems from Zeiss, Leica and MMI all use a UV laser as a kind of optical scalpel instead. According to Vega, this UV laser cuts in “the optical plane of the membrane” and ablates the tissue underneath the beam itself but spares both the cells of interest and those outside the circumscribed region. (As an added bonus, the UV laser can be used to ablate or destroy cells that are not of interest, notes Espina.)
Where these three systems differ is in details, such as how they perform the excision operation—Leica moves the laser, whereas Zeiss moves the stage—and how they collect the cells researchers want to study. MMI’s CellCut Plus uses an adhesive-coated collection cap that sticks to a membrane above the cells. Leica’s LMD7000 allows the cells to fall by gravity into a collection tube. And Zeiss’ PALM MicroBeam catapults the cells into the tube via a laser-induced pressure wave.
The MicroBeam’s catapulting force, explains Zeiss product marketing manager Timothy Pratt, is akin to “a light-induced micro-explosion into the matter lying just below your sample and the ensuing shock wave propelling the tissue upward.”
(The MicroBeam also has a direct contact mechanism that uses a “sticky cap” for cell areas too large for catapulting, Pratt adds. “That’s helpful, especially in proteomics, where you need bigger samples to get sufficient signals.”)
The Laser Capture Microdissection Core Laboratory at the East Carolina University Brody School of Medicine uses a Zeiss PALM MicroBeam. Lab co-director Barbara Muller-Borer says one key feature that figured in selecting the instrument was its ability to dissect live cells. (Today most, if not all, LCM systems can handle live cells, at least as an option.)
Muller-Borer uses the live-cell dissection capabilities in her work on cardiac development, where she separates (excises) differentiating mesenchymal stem cells from co-cultures with cardiac myocytes. Another user, Anagha Malur, uses the system to excise fixed lung granulomas and paired non-granulomatous tissue in a mouse model of sarcoidosis for comparative gene expression analysis via RT-PCR.
According to Malur, one critical aspect to successful LCM work is optimization—figuring out exactly how long a user can spend fixing, staining and microdissecting a sample before the biomarkers begin to go bad. In Malur’s case, she determined she had no more than 20 minutes, from when the slide was put into the instrument to moving the excised cells into lysis buffer, before the RNA began to degrade. “Each investigator has to check that out” for his or her own system, she says.
Advice for LCM neophytes
LCM is not an experiment in itself, says Vega; it is a way to prepare a sample for downstream applications. Inevitably, those downstream applications will be working with relatively small sample sizes, so sensitivity is key. Thus, optimization vis-à-vis sample size is essential when it comes to LCM.
Take reagents for downstream processing, for instance. Several companies offer RNA, DNA or protein isolation kits optimized for very small samples in general, or LCM samples in particular, including Qiagen, Ambion and Arcturus. “Typically we use Qiagen products,” Malur says, “but also Ambion.”
She adds, “You have to try out a couple kits before you say one works and another doesn’t, because even if something is what you’re used to, the researcher has to have an open mind with respect to [using a different product for] LCM.”
Espina suggests that users consider four key issues when preparing to do laser capture work. First, how was the tissue fixed? Is it formalin-fixed and paraffin-embedded (FFPE) or frozen? Is it fresh or from an archive? Do you know how it was prepared and handled, how long it sat and so on? Second, what do you hope to do with it? Or to put it another way, Espina says: “Which ‘omics technology are you going to use downstream?” DNA is relatively easy to collect, all agree, whereas RNA degrades rapidly. Third, how many cells will you need? Nucleic acids can be amplified and thus can be studied from relatively small numbers of cells or even single cells. Protein requires more material (from 10,000 to 50,000 cells, estimates Muller-Borer), because it cannot be amplified. And fourth, what extraction method will you use, and is it compatible with downstream experiments?
In other words, Espina says, planning is key.
Consider also how you will visualize the cells of interest. All LCM systems can handle basic H&E stained sections, but you can also highlight cells using fluorescent tags or even immunohistochemically—an approach called immuno-LCM, says Liotta.
At the Histotechnology Laboratory at SAIC-Frederick, a contract laboratory for the NCI, research associate Yelena Golubeva uses two systems for LCM work: an old PixCell IIe from Arcturus (IR laser only) and an MMI CellCut Plus.
The former, she says, is used for excision of relatively small regions, from a single cell up to regions of about 50 microns in diameter. For larger regions, she uses the MMI, as the laser cutting is much faster than on the PixCell IIe, and at SAIC-Frederick, LCM users pay by the hour.
According to Golubeva, every sample is unique when it comes to LCM. Mouse liver samples are relatively stable with respect to RNA, for instance, whereas prostate samples begin degrading almost immediately. “This is an empirical problem,” she says.
So, Golubeva recommends users start by running some pilot projects. Start with a set of slides from the tissue of interest. Test the stain you hope to use—does it damage the biomolecules you hope to study? Stain one section, scrape the cells off it and compare the RNA you extract from it with cells scraped from an unstained slide, both in terms of overall quality (for instance, on an Agilent Bioanalyzer chip) and in terms of qPCR data. “If that’s okay, then you can use this stain on other sections in the set,” she says.
Next, optimize timing. How long can you spend dissecting before the RNA quality falls off? Ten minutes? Twenty? And so on. “It doesn’t matter what machine you have, you need to know how the tissue … behaves with the LCM instrument and with the consumables that you use,” she explains.
Vega suggests a few other variables, such as cutting speed and power. These are functions of the laser itself, Vega explains. For instance, the Leica LMD7000 features a 120 microJoule laser firing at up to 5,000 Hz; the LMD6500 has a 70 microJoule laser firing at 80 Hz. Thus, the former can cut faster and handle thicker sections than the latter. The 7000 also can cut harder sections, he adds, tissues like bone, tooth and fibrous plant tissue that are traditionally tougher to cut than others.
One way to gain insight into those problems is to talk to other users. Muller-Borer and her co-director, Mary Jane Thomassen, are members of a North Carolina-based user group called LASER TAG, which meets four times a year to advance LCM technology and “provide a forum for users and vendors to discuss new applications and new research,” says Muller-Borer.
Most of the issues that come up at these meetings concern not the LCM process itself, says Thomassen, but the ancillary work of slide and sample preparation. Indeed, she says, laser capture microdissection is relatively easy. “It’s the preprocessing and postprocessing that’s difficult.”
But advances are being made on that front. Espina is also a member of LASER TAG, and over the years, says Liotta, she has trained “thousands of people” to use the technique. In Molecular Profiling: Methods in Molecular Biology, Espina described a new fixative to better preserve proteins and phosphoproteins for LCM work. And University of Connecticut researcher Joel Pachter recently published, in Journal of Neuroscience Methods, a streamlined method for transcriptome analysis that skips RNA extraction altogether, reducing the time from dissection to multiplexed qPCR analysis to seven hours.
With some 3,400 papers published to date on LCM, new applications and methods are coming online constantly. But for a good primer, check out Espina’s 2006 article in Nature Protocols, and her 2012 chapter on the ArcturusXT, also in Molecular Profiling: Methods in Molecular Biology.
The image at the top of the page is from Leica Microsystems’ LMD 6500 and LMD 7000.