by Jeffrey M. Perkel
When pathologists receive tissue specimens for microscopic examination, they have two options. They can encase the tissue in ice, or they can fix it with formalin and embed it in paraffin.
Both processes achieve the same goal—producing a solid tissue block from which thin slices can be shaved—and the latter option, called FFPE (Formalin-Fixed, Paraffin-Embedded), has been in use for nearly a century. Today there are millions upon millions of FFPE samples worldwide, many of them decades old and associated with detailed clinical data that makes them precious scientific resources.
There's just one problem: FFPE-treated samples work fine for immunohistochemistry and H&E staining, but for molecular methods like microarray analysis or quantitative PCR? Not so much.
As Ken Doyle, director of marketing at Epicentre Biotechnologies, explains, FFPE processing attacks macromolecules, especially nucleic acids, in several ways. First, formaldehyde is a cross-linking reagent that tethers nucleic acids to one another, as well as nearby proteins. "That is a chemical reaction and not easily reversible," he says. Cross-linked nucleic acids will not perform properly in most enzymatic processes, including PCR.
Nucleic acids can also be degraded and modified by the fixation process, and, independently, degrade with time. And of course, to extract from FFPE tissues, you must also remove (or at least melt) the paraffin.
In other words, those hoping to extract pristine genomic DNA or RNA from decades-old FFPE tissue blocks are out of luck. However, molecular tool manufacturers have developed a number of solutions to help extract nucleic acids (especially RNA) from FFPE blocks, and to ready it for downstream analyses.
Epicentre is one of a group of companies offering reagents for macromolecule preparation from FFPE samples. The company actually offers two types of products for this purpose, says Doyle: FFPE-ready QuickExtract kits for extraction of DNA or RNA, and a general-purpose MasterPure kit (which works with FFPE samples, but also with other types of inputs).
QuickExtract, Doyle explains, is "quick and dirty," but sufficient for PCR, a popular post-extraction option. "PCR is typically a little more forgiving than other techniques," Doyle says. "You can use fairly dirty material and still get results." MasterPure, on the other hand, requires a more involved protocol, including several precipitation steps, but produces cleaner product suitable for a variety of applications.
In a study published in December 20091, Hyung Kim of Cedars-Sinai Medical Center in Los Angeles, and colleagues—after testing five RNA extraction technologies on FFPE renal cell carcinoma samples dating from between 2002 and 2006—reported that MasterPure and Life Technologies' TRIzol reagent
performed best in terms of RNA yield. The other technologies tested were the RecoverAll™ Total Nucleic Acid Isolation Kit from Ambion , now part of Life Technologies; the High Pure FFPE RNA Isolation Kit from Roche Applied Science; and the ArrayGrade FFPE RNA Isolation Kit from SABiosciences , now part of Qiagen. (The authors also report that adding an overnight proteinase K digestion step to either protocol "resulted in a significant increase in RNA yield.")
Other FFPE sample prep options include Norgen Biotek's FFPE RNA and RNA/DNA purification kits, the Agencourt FormaPure system (available from Beckman Coulter), the Arcturus Bioscience Paradise Plus Reagent System (available from Molecular Devices, now part of MDS Analytical Technologies), and Life Technologies' Invitrogen PureLink FFPE RNA Isolation Kit. Qiagen supports perhaps the broadest range of applications, with kits for extraction of DNA (QIAamp DNA FFPE Tissue Kit), RNA (RNeasy FFPE Kit), microRNA (miRNEasy FFPE Kit), and protein (Qproteome FFPE Tissue Kit).
According to Janina Schaper, global product manager for diagnostic sample preparation at Qiagen's Hilden, Germany, facility, each of these kits can take as input either an FFPE slide or tiny tissue block of up to 25 mg, though the former is preferred. Tissue slices are used routinely for histomorphologic and microscopic examination, Schaper notes, "but also [facilitate] the sample lysis due to a bigger surface area accessible for the lysis solution," she says.
Generally, Schaper says, yields and quality from such processes vary dramatically from sample to sample and are highly dependent on sample age and fixation conditions. Doyle says 1cm2 of tissue typically produces 1-2 ug of DNA using the QuickExtract Kit.
Once you have extracted your macromolecules, you can analyze them—if you have enough material. Sometimes, however, there's too little material to do that directly. That's where amplification comes in. Qiagen's REPLI-g FFPE kit enables whole genome amplification of DNA from FFPE tissue, while NuGEN Technologies' WT-Ovation™ FFPE System V2 and Genisphere's RampUP and SenseAMP kits support RNA amplification.
RNA amplification requires a cDNA intermediate, and NuGEN's and Genisphere's kits are no exception, but with a twist: because the transcripts are often degraded, traditional oligo-dT-based cDNA synthesis won't work efficiently. As a result, both manufacturers employ both oligo-dT and random primers. That's where the similarity ends.
NuGEN's amplification process uses the company's Ribo-SPIA RNA amplification technology: a linear, isothermal approach that employs chimeric RNA-DNA primers to drive multiple cycles of DNA polymerase-driven synthesis, resulting in an antisense cDNA pool. Genisphere creates an RNA pool by coupling a T7 promoter to the 3' end of the first-strand cDNA and using that to drive sense-strand RNA synthesis.
Robert Getts, Genisphere's chief scientific officer, coauthored a recent study comparing various RNA extraction and amplification methods for FFPE tissues.2 The authors achieved their best results using Ambion's RecoverAll and Agencourt's FormaPure extraction systems, but other kits will also work, Getts says—just be sure to use the same method throughout the experiment. "You cannot use RNA purified by one kit and expect to get good results when comparing to RNA extracted with another kit." That's because each method "tended to have its own bias towards how RNAs were extracted," Getts explains.
When selecting a kit, Getts advises testing the resulting gene expression profile against fresh frozen materials. "Users should ask, given the tissue source I am working with, which methodology gives me the appropriate signature for that tissue that I would expect if I were processing it using a frozen methodology?" That way, he says, "they will see the least data loss."
The sample will still be degraded, however—a factor you must bear in mind when planning downstream experiments. In particular, says Iain Russell, senior product manager for TaqMan® MicroRNA Assays at Life Technologies, when designing qPCR assays for mRNA expression, think small. Life Technologies allows users to filter its TaqMan assay selection software by amplicon size (e.g., less than 70, 71-85, 86-100, or greater than 100 bases).
"By biasing the amplicon size to shorter amplicons, that opens up the ability to use these assays with lower-quality samples, and FFPE samples fit into that category," Russell says.
Alternatively, consider microRNAs, which are relatively unaffected by FFPE treatment, says Russell, probably because of their small size.
"MicroRNA has really been an area where a lot of our users have seen what they would consider to be considerable success," Russell says. "I see a lot of studies moving forward using FFPE samples looking at microRNA biomarkers as a result of that."
Life Technologies offers microRNA assays both individually and in sets of up to 381 assays on microfluidic cards.
Of course, users can always opt to outsource their FFPE gene expression analysis instead. Exiqon, which focuses on microRNA, can extract clients' nucleic acids from FFPE slides or tissue blocks and profile their expression using microarrays, qPCR, or in situ hybridization, according to Chris Harbert, the company's director of marketing.
For many researchers, the precious nature of FFPE samples, combined with the difficulty in working with them, makes outsourcing them a sensible alternative, says Harbert.
"You can't go back to the patient and get a new biopsy," he says.
1Glenn, ST, et al., "Maximizing RNA yield from archival renal tumors and optimizing gene expression analysis," J Biomol Screen, 15:80-5, 2010.
2Roberts, L, et al., "Identification of methods for use of formalin-fixed, paraffin-embedded tissue samples in RNA expression profiling," Genomics, 94:341-8, 2009.
Main article image is from NuGen.