July 2011
Feature Story

Anne Paxton

A new system from Siemens Health-care Diagnostics, developed with the help of nanotechnology, could strengthen molecular diagnostics through its fully automated, walkaway method for extracting nucleic acids from his­to­path­o­log­ical tissue samples.

Slated for worldwide launch later this summer, the system, called the Siemens Tissue Preparation Solution, offers simultaneous extraction of RNA and DNA from a single formalin-fixed, paraffin-embedded sample, and promises a throughput of up to 48 samples in about four hours. By employing paramagnetic particles with a nanolayer of silica, the system is able to de-paraffinize the samples and purify the nucleic acids. It is designed to support routine clinical diagnostic analysis in the molecular pathology laboratory, as well as extensive retrospective biomarker studies of archived FFPE tissue material.

“It’s a really unique and complete automation system including de-paraffinization, and it delivers high-quality nucleic acids, both RNA and DNA,” says inventor Guido Hennig, PhD, senior global scientific affairs manager for Siemens Healthcare Diagnostics in ­Esch­born, Germany. “It can be applied both in the research ­arena to exploit the FFPE archives, and also for clinical molecular applications. In principle, I think Siemens has set a new standard in extraction of RNA and DNA from FFPE samples for any downstream application.”

The instrument and reagents are IVD labeled and CE marked for European distribution, and they will be class 1 exempt in the U.S. when listed with the FDA and will not require a clinical trial or 510(k) application, says Ellen Sampson, MS, MBA, senior global marketing manager, Siemens Healthcare Diagnostics, Tarrytown, NY, who is in charge of the launch. The system is geared to mid- and high-volume laboratories.

Like a spaceship making the jump to light speed, fully automated DNA/RNA extraction could sharply accelerate the introduction of biomarkers using extracted nucleic acids from histopathological tissue specimens into routine clinical practice. Eliminating the manual steps in this process will allow researchers to efficiently plumb the huge amount of data locked in stored tissue samples up to 30 years old. DNA and RNA could formerly be extracted from these archived samples only manually or with partial automation. According to Siemens, the Tissue Preparation Solution’s success rate in extracting PCR detectable RNA or DNA from several thousand FFPE samples of up to 30 years of age has been shown to be about 99 percent.

The data that will now be available could produce a tantalizing array of detailed answers to the general question of which genetic profiles combined with which treatments lead to which patient outcomes. Will this speed the development of targeted diagnostics? Absolutely, Sampson says. “You’ve got well-characterized, aged samples. You know what ultimately has happened to the patient. It now becomes more valuable to look at that tissue and how the DNA and RNA relate to the outcomes.”

“There’s basically a gold mine of data that’s been untapped because of the difficulty of extracting samples from the FFPE,” says Toumy Guettouche, PhD, director of new genome technology assessment and implementation, Hussman Institute for Human Genomics, and director of the Oncogenomics Core Facility, Sylvester Comprehensive Cancer Center, University of Miami School of Medicine. He is a pilot tester of the Siemens system.

As Dr. Hennig explains, medical researchers have concentrated their efforts on expression profiling, genotyping, and mutation analysis to provide guidance in the treatment of cancer. But the number of individual markers or panels of markers using extracted nucleic acids introduced into routine clinical practice has been limited. With fully automated extraction, that limitation could fall. “There are hundreds of millions of these FFPE samples in archives all over the world, many with followup and outcomes data,” Dr. Hennig says. “Both high- throughput and fully automated solutions will help us cover this large volume of samples.”

The most laborious parts of manual and semiautomated extraction are de-paraffinization and lysis, says Dr. Hennig. “De-paraffinization can take significant time, and manual methods use variable lysis incubation times, often requiring an overnight digestion step. The specialized hardware we’ve developed allows a unique fully automated workflow including lysis and de-paraffinization.” Most protocols for getting rid of paraffin require the use of xylene and ethanol, which are hazardous to use and to dispose of, he notes. “We make that obsolete. De-paraffinization is done by a very elegant, safe, hydrophobic adsorption step. And we standardize our lysis time to only one hour, independent of tissue type to be extracted.”

Nanotechnology is the innovation that makes it possible to automate de-paraffinization—in this case, magnetic particles with bead sizes less than 1 μm. “All other devices for extraction of nucleic acids that employ magnetic particle technology have a high content of silica. We have proprietary paramagnetic particles with a nanolayer of silica adsorbed to the iron oxide particle core. The semiautomated systems also use magnetic particles, but not nanobeads.”

What is proprietary in the system is not only the beads but also the hardware to facilitate de-paraffinization, Sampson says. FFPE sections are loaded directly onto the Siemens platform in 1.5-mL plastic tubes. “We automate the lysis step and the heating and cooling step, which first of all breaks apart tissues so that nucleic acids are released in solution. You have a lot of tissue debris in the solution, and when the paraffin is cooled, it forms a ring around the inner wall of the tube. That leaves a hole so you can pipette out your supernatant fluid.”

The tissue debris, however, leads to clogging. “You have to remove the tissue debris so you can pipette the fluid out, and a lot of other semiautomated systems would need to do an offline centrifugation step. But we don’t require any centrifugation or offline processing. Instead, we take our magnetic beads—and they have different properties depending on what buffer is being used along with them—and we have nonselective or negative binding, where the tissue debris binds to the beads’ silica surface. Then you can magnetically pull the beads and tissue away from the supernatant fluid and aspirate it out into a different vessel.”

In the second step, a fresh aliquot of the same magnetic particles selectively binds to the nucleic acids. “So we can use the same beads to first nonselectively pull off tissue debris, then use the beads under different buffer conditions—called chao­tropic buffer solution—to selectively bind the nucleic acids.”

Siemens became involved in developing this automated nucleic acid isolation system through its genomics research program, Sampson says. “This is a technology we used to complement our internal research. In talking to pathologists and researchers around the world, we realized there was a huge opportunity here, and we don’t believe anybody has a fully automated system yet available, so there was a window of opportunity for us.”

To speed the product’s launch, Siemens focused on one sample type, she explains. “FFPE is probably the sample type that’s most commonly acquired in clinical practice, and we felt it was one of the most difficult samples to extract nucleic acid from, so we decided to focus on that first.”

“Manual extraction from FFPE samples is not amenable to high through­put, and we have been looking for an automated solution for some time,” says Dr. Guettouche. His combined cancer center/Hussman Institute core facility at the University of Miami has been a lead user, or tester, of Siemens Tissue Preparation Solution for about six months. “I was working for Bayer Diagnostics [Siemens acquired Bayer Diagnostics in 2006] when they were testing those beads. So I knew about the physical bead properties and chemistry, and I contacted Guido Hennig to ask how it was developing. They told me they had automated the chemistry and were developing it for FFPE, and we were very interested and struck up a collaboration.” Dr. Guettouche’s goal was to automate pro­cesses to make them more cost-effective, to increase throughput, and to reduce variations.

As part of its research program, his facility has been testing how reliable and reproducible the instrument is, and it has extracted RNA and DNA from two different tissues so far, breast and prostate. “We’re very pleased with the results. These magnetic beads are fantastic. It’s a very innovative step to get rid of paraffin using hydrophobic adsorption. The quality and quantity of the extracted nucleic acids are comparable to or better than the manual methods we were using before—and we were arguably using one of the best methods to extract.” One advantage of the Siemens system is that “the fragments we’re getting are on average larger than other commercial methods’ fragments.” The laboratory is now testing five other tissues: lung, pancreas, kidney, bladder, and cervix. At this stage, “we’re not looking for particular markers. We just want to see if we can reliably extract RNA and DNA from all of these tumor tissues.”

Dr. Guettouche and colleagues are already starting the next phase: “We’re actually going to applications possibly leading to clinical interventions. We’re looking mainly at next-generation sequencing so we’ll be doing targeted resequencing, whole transcriptome, and whole genome sequencing.” Among other things, “we want to do targeted resequencing and whole transcriptome studies of tissues that are in our biorepositories, then look for biomarkers. If you want to analyze some tumors that we have here, and see what type of mutations they have and see what the treatments or drugs and outcomes were, you can look at which patients survived and which are deceased. And you can try to do a prediction, basically developing a classifier for tumors. Then you can perform more targeted treatments.”

The center is also beginning a project to extract nucleic acids in very small quantities from laser captured microdissected samples. In theory, the system can be used on all kinds of sample input formats, including tiny amounts of tissues such as tissue microarray cores, fixed sections from needle biopsies, snap frozen tissue, and cells from culture.

He is working now on several articles to demonstrate the utility of high-throughput systems for FFPE extraction. However, one of the problem areas is quality control of the material that comes out of the system. “There’s not really any agreement on what quality control you need to assess the material that has been extracted from FFPEs. We’re also working on figuring out what are the changes at the nucleic acid level that happen when you do formalin fixation and how they affect sequencing results. So we are establishing quality control parameters for downstream applications like next-generation sequencing, and we’re looking at developing applications for cancer and complex genetic diseases like Parkinson’s or Alzheimer’s, where we can analyze all the samples in the archives.”

Hussman Institute/Sylvester will also conduct research to compare fresh frozen specimens with FFPE specimens. “With all sample types, it’s very important that you have a standardized protocol for how to freeze or fix them. What we’ve found is that the quality of samples you get out of FFPE or fresh frozen varies widely. Part of that is how they process the sample right after they take samples from the tumor or biopsy. Did they immediately freeze or fix it in formalin? Did they use buffered formalin? How long did they fix it? And if it’s fresh frozen, did they use a preservative like RNA later? The protocol will influence the quality of nucleic acids you extract. So it’s not only the extraction method but the method that was used to preserve the tissue or sample.”

For example, if the surgeon takes out a piece of tissue and puts it in a dish, “to preserve it you have to immediately put it in formalin or liquid nitrogen or whatever solution you’re going to use. If you let it sit at room temperature, the nucleic acids degrade really rapidly, especially the RNA. The longer it sits, the worse the quality of nucleic acid is going to be. And if they exceed the optimal time of fixation, it’s also difficult to get good material out of it. So there are lots of parameters playing a role in the quality of extracted nucleic acids.”

There’s not that much in the research literature on deep sequencing of FFPEs, he notes. “There are fewer than 10 papers published on this so far.” At the University of Miami, “we used to have several different biorepositories all running different protocols. Everybody was doing their own thing. But now the Hussman Institute and the pathology department have taken the lead and implemented standardized procedures for how to generate and process FFPEs and fresh frozen tissue.”

Another lead user of Siemens’ automated purification system, the University Hospital Charité Berlin in Germany, has been working with it for more than a year. “Siemens has been discussing with us for years what new technologies are needed, and one was an instrument that might help us extract RNA and DNA,” says Manfred Dietel, MD, director of the hospital’s Institute of Pathology.

His laboratory has been conducting most of its extractions on a semiautomated platform. “But due to the fact that molecular diagnostic pathology is much more involved in the decision whether patients will receive this or that drug, it becomes more and more important for us to standardize all our technologies, starting with the fixation process, then extraction, and then the sequencing.” Given the 4,000 to 5,000 clinical cases his laboratory processes each year, and several research projects, the semiautomated system is still in operation, “but when the Siemens system goes on the market, I think we will get another one or two more of those.”

The instrument’s walkaway automation is an advantage. “It’s very solid and very robust, and we’ve never had any really big problems with it. Once the technician understands it, she loads it and it’s done four hours later, and in between she can work on other tasks.”

To perform molecular analyses, “we need to extract DNA and RNA from FFPE every day,” Dr. Dietel notes. Over the past five years, he estimates, there has been a threefold or fourfold increase in the extraction of nucleic acids for use in molecular diagnostics in Europe.

Identifying and validating molecular markers in FFPE tissue is an area of intense and exciting research activity, Dr. Hennig says. Through the platform’s high-throughput testing of biomarkers in routine clinical diagnostics and retrospective and pro­spective research studies, he believes Siemens’ nucleic acid isolation technology will make a significant contribution to patient care.