FFPE and RNA Extraction

Table of Contents

Letter from the Editor
Index of Experts
Q1: How do the age and tissue storage conditions affect your choice of RNA extraction method?
Q2: How do you optimize your extraction method for the best purity, yield, and fragment size of RNA?
Q3: How do you check the quality of RNA extracted, and what determines your choice?
Q4: How do you tell if the RNA is suitable for expression profiling (i .e ., the level of RNA degradation)?
Q5: What method do you use to measure transcript expression (microarray, RT-PCR), and what determines your choice?
Q6: How do you validate and/or normalize your expression data?
List of Resources

Download the PDF version here

Letter from the Editor

In this month's issue of GT, we bring you a technical guide on FFPE and RNA extraction. While fresh or frozen samples are ideal when it comes to RNA quality, RNA from formalin-fixed, paraffin-embedded (FFPE) tissue allows for gene expression studies on archival samples. Older samples are especially promising for high-throughput research in that they're typically linked to long-term phenotypic data and can be used for clinical analyses in, for instance, large population-based 'omic studies.

The main problem with FFPE samples is that the RNA has typically been degraded, so getting the most out of the de-paraffinization and extraction process is the trick to reliable and reproducible gene expression profiling, whether you're running microarrays or RT-PCR. And while a variety of kits exist for extracting RNA from FFPE tissue, it's how you tweak the protocol that will yield the best results.

In this guide, there are questions ranging from how to optimize protocols so that the RNA is suitable for gene expression studies to how to get the most accurate results in the shortest possible time. Whether you're aiming for purity, yield, or best fragment size, or aren't sure if your RNA is suitable for your chosen profiling assay, our experts cover all the bases. For additional help, turn to the resources guide at the back. Good luck extracting!

— Jeanene Swanson

Index of Experts

Many thanks to our tireless experts for taking the time to contribute to this technical guide, which would not be possible without them.


Falko Fend
University of Tübingen

Stephen Hewitt

NIH/NCI

Rolf Jaggi
University of Bern

Beatrice Knudsen

Fred Hutchinson
Cancer Research Center

Jun Luo

Johns Hopkins Hospital

Alfredo Ribeiro-Silva

University of São Paulo

Q1: How do the age and tissue storage conditions affect your choice of RNA extraction method?

Dealing with archival FFPE tissues, time of fixation, and the use of buffered versus non-buffered formalin as well as the age of the paraffin blocks are the main determinants of RNA quality. In contrast, pre-fixation time seems to be less important for many mRNA species than previously thought, but this may strongly depend on the transcripts analyzed and ambient conditions. Especially for older blocks or samples from other institutions, these parameters are often unknown or poorly controlled. For this reason, we use the same extraction technique for all FFPE samples, including microdissected tissue sections. A critical point is sufficient digestion time with proteinase K at high temperature, i.e. overnight at 60°C. Shorter times and lower temperatures reduce the yield of RNA from FFPE. We use a more time-consuming method based on extraction with phenol/chloroform (Specht K, et al., 2001; Koch I, et al., 2006) with excellent yields, but RNA extraction kits specially designed for FFPE samples have shown good results, such as the FFPE kit from Ambion. Proper DNase I digestion is mandatory unless only intron-spanning primer sets are used, since DNA contamination may provide a significant source of error. However, even with an optimal protocol, old samples will sometimes yield less amplifiable RNA.

— Falko Fend

It is clear that older specimens, or those stored in less than optimal environments contain less/lower quality RNA, but we do not use these factors in determining the extraction method. We approach RNA from FFPE as 'more is always better.' The bigger question is, why is RNA yield/quality reduced in older specimens? We are trying to determine the factors that might control this, both endogenous to the specimen, and exogenous in the storage environment. We have more information on slides than blocks, but it is safe to say both are factors. We know that differences in tissue processing directly affects RNA quality and quantity. What we don't know is if these are the underlying cause of poor RNA recovery from older blocks, or if this contributes to RNA degradation. Stay tuned for details, but I think it is safe to say heat and humidity are bad.

— Stephen Hewitt

We observe a slow but consistent decrease of qRT-PCR efficiency with storage time of FFPE blocks. The effect is gene-independent. We currently use a proprietary protocol to reduce chemical modifications in RNA which occur during fixation with formalin. A kit containing all components and a detailed protocol is available from AmpTec, Hamburg, Germany (http://www.amp-tec.com). Our protocol was successfully applied with FFPE material that was stored for up to 8 years.

— Rolf Jaggi

Age and tissue storage don't appear to affect RNA quality.

— Beatrice Knudsen

In general, FFPE tissues that have been stored at room temperature for a prolonged period of time (e.g., years) pose greater technical challenge than those fixed and embedded recently (e.g., months ago) or stored frozen following FFPE. Most methods employ an initial proteinase K digestion step. For recently processed FFPE tissue, we generally prefer an overnight incubation with the proteinase K mix at 50°C. For those more challenging FFPE tissues, we would elevate the incubation temperature and prolong the incubation time.

— Jun Luo

The major advantage of extracting RNA from formalin-fixed paraffin-embedded tissues is the possibility of gene expression studies on archival samples. For that purpose, older samples are particularly important because they are associated with longer clinical follow-up. However, according to literature, RNA from tissues archived for several years is more degraded than those ones archived for less than one year, and consequently less suitable for molecular tests. Curiously, in our experience, we did not find a significant difference in quality from the RNA extracted from breast samples stored for 10 years and breast samples stored for less than year. The samples were stored at room temperature. The differences between the older and the newer samples relied on the quantity, consistency, and success rate of extracting RNA, but not quality. The quality was measured by the RNA integrity number, using the Agilent 2100.

— Alfredo Ribeiro-Silva

Q2: How do you optimize your extraction method for the best purity, yield, and fragment size of RNA?

Using quantitative RT-PCR assays for a small set of reliable housekeeping genes such as TATA box binding protein, the extraction method can be fine-tuned for optimal RNA yield and quality. When comparing RNA from FFPE and frozen tissue, the average shift in real-time RT-PCR is in the range of 3-4 Ct. In general, average mRNA fragment size from FFPE is well below 500 bp, and amplicon sizes of quantitative RT-PCR assays should therefore be below 100-120 bp. of note, the Ct shift between fresh frozen and FFPE derived RNA for a given assay even in the optimal size range (<120 bp) may vary from less than 2 to 5 Ct, although this Ct shift remains constant for a given assay. These differences have to be taken into account if RNA transcript levels are directly compared, such as splice variants of a gene.

— Falko Fend

We use a full court press approach focusing on chemical factors in extraction. We have not found a set of conditions that favors purity over yield or fragment length. We found that the greatest failing of most extraction protocols was the de-waxing procedure. I am not convinced our de-waxing is complete, but we can demonstrate that longer times or higher temperatures would yield more/better RNA. Our paper was the first to quantitatively address RNA yield in mg-to-mg comparison of frozen and FFPE tissue.

— Stephen Hewitt

We tested various column-based protocols from commercial providers and compared them to our own column-based protocol. Due to unsatisfactory results with these kits we developed our own protocol and reagents for RNA isolation from FFPE material. The reagents work well with a broad variety of FFPE samples (e.g., breast, prostate, and lung cancer; normal breast, skin, liver, muscle, brain, and lung).

— Rolf Jaggi

The best methods are those that don't require RNA extraction. They work by annealing of probes to RNA that is cross-linked in the tissue. A proteinase K digest liberates the RNA from the tissue, but the RNA is not isolated. RNA isolation generates a bias for certain genes and the bias needs to be evaluated on a gene per gene basis.

— Beatrice Knudsen

We optimize the extraction method mainly by varying the conditions at the initial proteinase K digestion step, e.g., enzyme concentration, temperature, incubation time. To facilitate complete digestion, we always use microtome tissue sections prepared under RNase-free conditions.

— Jun Luo

We tried several protocols comprising two different methodologies. In one of these methodologies the RNA is extracted in a spin column of purification, and in the other one, the extraction is magnetic bead-based. One particularly important step is the de-paraffinization of the sections obtained from the paraffin blocks. In our experience, among the column of purification protocols, we got better results using the de-paraffinization solution dlimonene from Stratagene instead of xylene. However, we liked more the results obtained by the magnetic bead-based protocol in which the de-paraffinization is made using the lysis buffer provided by the supplier. We used the Agencourt formaPure kit. This kit utilizes the solid Phase Reversible Immobilization (sPRI) paramagnetic bead-based technology to isolate RNA. Besides being easier and less time-consuming than the column-of purification-based protocols we have tested, this kit also extracted more RNA.

— Alfredo Ribeiro-Silva

Q3: How do you check the quality of RNA extracted, and what determines your choice?

We use the Nanodrop spectrophotometer for determination of RNA quality and purity, although Ct values obtained in subsequent qRT-PCR can still vary considerably depending on the individual sample, if the same amount of input RNA is used.

— Falko Fend

We use a multi-step approach, starting with a Nanodrop reading and 260/280 ratio . from there, it goes onto a Bioanalyzer RNA chip from Agilent. Our greatest concern is not consuming the precious RNA in the quality testing, but we do not want to waste reagents and time on poor-quality RNA. In general, when starting a project we will beta test the material, choosing what we expect to be the best and worst material, or randomly, to give us some general idea of quality. We won't skip the Nanodrop, but for low cellularity specimens, or microdissected specimens, the Bioanalyzer is not practical. This is a good example where we might assay the RNA quality on a whole section, but then isolate the RNA from cells of interest after some form of dissection.

— Stephen Hewitt

We routinely check the quantity and quality of the RNA by Nanodrop (E260, E260/E280 ratio, spectrum 220-320 nm) and by separation on an Agilent Bioanalyzer. Especially the spectrum may indicate that residual salt from the lysis buffer contaminates the RNA. This may inhibit downstream applications like reverse transcription.

— Rolf Jaggi

If we expect long RNA fragments, we use the Bioanalyzer. If the RNA comes from FFPE tissues, we purify some and use actin primers of different distance between the primers in a qPCR. The 3' primer is very close to the polyA and the three 5' primers are at increasing distance in the 5' direction. Each primer set will provide a signal intensity and we calculate the ratio between the primer sets to come up with an RNA quality index. The dynamic ranges of measuring RNA fragment length differs between the Bioanalyzer and the qPCR method and there is no overlap in the range.

— Beatrice Knudsen

We mainly rely on assessing the electropherogram from the Agilent Bioanalyzer. The most important indicator of RNA quality, of course, is the presence of the 28s and 18s ribosomal bands. However, even in the absence of ribosomal bands, the extracted RNA could also be useful in a number of applications, though we would try to put them on expression arrays. In this case, the shape of the plateau following the small RNA peak provides indications regarding the relatively RNA quality. If they taper off gradually towards high MW RNA, RNA is relatively good. A sharp dip off the small RNA peak is an indicator for bad quality.

— Jun Luo

We check the quantity and quality of RNA using the Agilent 2100 Bioanalyzer. The Bioanalyzer utilizes a combination of microfluidics, capillary electrophoresis, and fluorimetry to determine RNA length, distribution and concentration. The quality is expressed by the RNA integrity number (RIN). Because in stored paraffin blocks there is a tiny quantity of degraded RNA we use the RNA 6000 Pico kit from Agilent rather than the Nano kit. The quantity is given by pg/µl.

— Alfredo Ribeiro-Silva

Q4: How do you tell if the RNA is suitable for expression profiling (i.e., the level of RNA degradation)?

You have to bite the bullet and perform the assay. We use internal genes, aiming for somewhere around 10 percent of transcripts as controls. The more targets we interrogate, the more controls we can utilize. We wish we could do better, but we have no magic RIN number. If you have seen the Bioanalyzer tracings, you realize, you cannot calculate a meaningful RIN.

— Stephen Hewitt

We established qualitative and quantitative parameters to judge the RNA quality. based on the distribution of RNA fragments on a Bioanalyzer we can identify samples with poor RNA quality which are not suitable for qRT-PCR applications (low efficiency — elevated Ct values). Apart from this we perform qRT-PCR experiments with a series of three to five reference genes with constant amounts of RNA. The mean of several Cts must be lower than 30, otherwise the sample is treated as unsuitable for qRT-PCR. several reference genes must be tested on a larger series of samples and optimal genes must be identified separately for each tissue.

— Rolf Jaggi

We use the Quantigene 2.0 platform for measurement of RNA in FFPE tissues and we use their threshold recommendations for the cutoff of suitability. The threshold was generated by measurements of ribosomal DNA and RNA.

— Beatrice Knudsen

We look for the ribosomal bands. If we find evidence for the presence of ribosomal bands, even at trace amounts, we would pursue the samples for expression profiling. We have generated good quality expression microarray data from RNA samples with RIN number at 2 .0.

— Jun Luo

We use an arbitrary RIN of 1 .4 as a threshold for a minimally acceptable quality for gene expression analysis. This is very low RIN (the maximum value is 10); however, using primers targeting 151 bp fragments of the ubiquitous gene glucose-6-phosphate dehydrogenase (6gPd), we were able to extract RNA suitable for RT-PCR in all cases stored for 10 years in which the RIN was higher or equal to 1 .4. Using primers targeting 242 bp fragments the successful rate of extracting biological useful RNA dropped to 43 percent. In that way, the length of RNA template seems to be more important than the RIN itself.

— Alfredo Ribeiro-Silva

Q5: What method do you use to measure transcript expression (microarray, RT-PCR), and what determines your choice?

We generally use quantitative RT-PCR, which is a reliable and robust technique also suitable for very small amounts of tissue, such as purified cell populations obtained by laser microdissection. For quantification, we prefer TaqMan assays with specific probes for each transcript, since intercalating dyes such as SYBR green are not reliable enough for FFPE samples.

— Falko Fend

It all depends on the goals of the project. We have used microarrays, and have obtained very nice results, but it is expensive and the difference in RNA quality can swamp meaningful results. In the current environment I would favor that approach for discovery. RTPCR is great when the number of genes is very limited and will be assayed on a regular basis. We have used the branched DNA technologies (Quantigene Assay, Panomics). They seem to fill a real niche. Well based, you can examine two or three genes without much assay development, but add a Luminex, and it is easy to scale up to a reasonably sized panel using the same probe designs. What we like about this approach is the lack of enzymatic steps — isolate RNA and direct assay — no reverse transcriptase. We have found you can get RNA, but the RNA may have so many adducts from cross-linking that reverse transcriptase is very unable to synthesize a cDNA. The branch DNA assays only require base-pairing. That said, we are moving toward deep sequencing.

— Stephen Hewitt

We regularly measure gene expression by qRT-PCR using a one-step protocol. We also used RNA from FFPE material for DNA chip experiments. Instead of using an oligo(dT)-T7 primer for the initial RT reaction, it is important to use a random primer or the unique TR inucleotide primer (AmpTec, Hamburg, Germany).

— Rolf Jaggi

We do this on a small scale and use PCR because it is the cheapest and most convenient for us and because we are used to it.

— Beatrice Knudsen

We mainly use expression microarray to measure transcript expression. We would pursue microarray analysis if we see evidence for the presence of ribosomal bands. For RT-PCR, the requirement is not as stringent, especially for those abundant transcripts.

— Jun Luo

Using RNA extracted from FFPE samples, we only got satisfactory results with RT-PCR technique. We tried to perform cDNA microarray but it did not work even with the highest RINs we were able to extract from FFPE samples. Our highest RIN was 6.1 (most cases had RIN between 2.0 and 2.4).

— Alfredo Ribeiro-Silva

Q6: How do you validate and/or normalize your expression data?

Using standard curves from serially diluted external plasmid standards, we determine in each sample the copy numbers of the target gene and of one or several well-established housekeeping genes. The resulting copy numbers are used to calculate the ratio of expression of the target gene in comparison to the housekeeping gene(s), since absolute copy numbers cannot be used due to the variable RNA degradation in FFPE tissues. The efficiencies of the assays in archival FFPE tissues are assessed beforehand, using matched frozen and FFPE samples (Koch I, et al., 2006). Using this approach, reliable and reproducible measurements can also be obtained from small amounts of FFPE-derived RNA, as long as the level of expression of the target gene is not in the very low range.

— Falko Fend

Validation depends on the goals of the experiments. We tend to focus on the development of assays for the clinical environment rather than discovery. "Validation" becomes verification (determining the observation is true) and validation (determining the test has the designed utility). It is a multistep process. Verification should include testing the limits of the assay, such as how little RNA is required. For validation, you want an independent cohort. Ultimately you want to demonstrate that another lab can apply your test on their samples and get the same result.

— Stephen Hewitt

Raw Ct values from qRT-PCR are normalized by calculating relative expression. For our breast cancer study we determined a set of three stably expressed reference genes from a set of nine putative reference genes. They were selected from analyzing a panel of about 80 RNAs derived of independent FFPE tumor samples. For dNA chip-based data we used quantile normalization.

— Rolf Jaggi

We use qPCR to validate expression data.

— Beatrice Knudsen

We validate the accuracy of gene expression data from FFPE tissue through comparison with those generated from the golden standard frozen tissues. Overall 80 percent concordance can be achieved if expression data is derived from FFPE RNA.

— Jun Luo

The best way to validate the data from FFPE samples is to compare the results with frozen tissue of matched samples, which is considered to be the golden standard for molecular procedures.

—Alfredo Ribeiro-Silva

List of Resources

Our panel of experts referred to a number of publications and online tools that may be able to help you get a handle on interacting proteins. Whether you're a novice or pro at the FFPE and RNA extraction game, these resources are sure to come in handy.

Publications


Abramovitz M, Ordanic-Kodani M, Wang Y, Li Z, Catzavelos C, Bouzyk M, Sledge gW Jr, Moreno s, Leyland-Jones B. Optimization of RNA extraction from FFPE tissues for expression profiling in the DASL assay. Biotechniques. 2008 Mar;44(3):417-23.

Chung JY, Braunschweig T, Hewitt SM. Optimization of recovery of RNA from formalin-fixed, paraffin-embedded tissue. Diagn Mol Pathol. 2006 Dec;15(4):229-36.

Chung JY, Braunschweig T, Williams R, Guerrero N, Hoffmann KM, Kwon M, Song YK, Libutti SK, Hewitt SM. Factors in tissue handling and processing that impact RnA obtained from formalin-fixed, paraffin-embedded tissue. J Histochem Cytochem. 2008 Nov;56(11):1033-42.

Cronin M, Pho M, Dutta D, Stephans JC, Shak S, Kiefer MC, Esteban JM, Baker JB. Measurement of gene expression in archival paraffin-embedded tissues: development and performance of a 92-gene reverse transcriptase-polymerase chain reaction assay. Am J Pathol. 2004 Jan;164(1):35-42.

Hewitt SM, Lewis FA, Cao Y, Conrad RC, Cronin M, Danenberg KD, Goralski TJ, Langmore JP, Raja RG, Williams PM, Palma JF, Warrington JA. Tissue handling and specimen preparation in surgical pathology: issues concerning the recovery of nucleic acids from formalin-fixed, paraffin-embedded tissue. Arch Pathol Lab Med. 2008 Dec;132(12):1929-35.

Mckinney MD, Moon SJ, Kulesh DA, Larsen T, Schoepp RJ. Detection of viral RNA from paraffin-embedded tissues after prolonged formalin fixation. J Clin Virol. 2009 Jan;44(1):39-42.

Paik S, Kim CY, Song YK, Kim WS. Technology Insight: Application of molecular techniques to formalin-fixed paraffin-embedded tissues from breast cancer. Nat Clin Pract Oncol. 2005 May;2(5):246-54.

Penland SK, Keku To, Torrice C, He X, Krishnamurthy J, Hoadley KA, Woosley JT, Thomas NE, et al. RNA expression analysis of formalin-fixed paraffin-embedded tumors. Lab Invest. 2007;87:383-391.

Ribeiro-Silva A, Zhang H, Jeffrey SS. RNA extraction from ten year old formalin-fixed paraffin-embedded breast cancer samples: a comparison of column purification and magnetic bead-based technologies. BMC Mol Biol. 2007 Dec 21;8:118.

Rupp GM, Locker J. Purification and analysis of RNA from paraffin-embedded tissues. Biotechniques. 1988 Jan;6(1):56-60.

Schroeder A, Mueller O, Stocker S, Salowsky R, Leiber M, Gassmann M, Lightfoot s, Menzel W, Granzow M, Ragg T. The RIN: an RNA integrity number for assigning integrity values to RNA measurements. BMC Mol Biol. 2006 Jan 31;7:3.

Conferences

Chi 9th Annual Genomic Sample Prep: Quality from the Ground up
http://www.healthtech.com/gta/gtl/spg