Feature Story

Everyday DNA—extraction gets easier

The special needs of RNA processing

March 2001
William Check, PhD

"Scope mouthwash—proven in scientific tests to be the best buccal cell collection medium for obtaining high-quality, high-yield DNA for clinical and research applications, including Southern blotting, amplification analysis, sequencing, and archiving."

No, you’re not likely to see that claim in an advertisement. But it is true. Researchers at Gentra Systems, a maker of kits and instruments for nucleic acid extraction, harvested buccal cells using several mouthwashes and saline and compared the DNA obtained from the cells. The verdict: DNA isolated from buccal cells was just as good as DNA from whole blood. And Scope was the best collection medium tested (Arch Pathol Lab Med. 2001;125:127-133).

"We were trying to expand our work with noninvasive methods for nucleic acid collection," says Ellen Heath, PhD, Gentra’s vice president of chemistry R&D. "Buccal cell sampling will be used more widely in clinical genetics," she notes, particularly in pediatric populations.

While Gentra’s definition of an optimal collection medium for buccal sampling may not qualify as a scientific breakthrough, it does illustrate important points about the field of molecular diagnostics. First, new applications of nucleic acid analysis arise very fast. Second, commercial or hospital-based practitioners of this discipline, many of whom have training in the basic methods of recombinant DNA technology, often assemble new methods from readily available materials.

Most important, ad hoc, in-house, manual methods of nucleic acid extraction and preparation continue to predominate much longer than anyone would have wished. Paradoxically, routinization and automation have been slow in coming to this highest of high-tech fields. But it looks like they may finally be arriving.

"Making nucleic acid, either DNA or RNA, is obviously something that we have to deal with every day in the clinical molecular laboratory," says Karen Kaul, MD, PhD, director of the molecular diagnostics laboratory at Evanston (Ill.) Northwestern Healthcare and associate professor of pathology, Northwestern University School of Medicine. Nucleic acid may need to be extracted from human tissues or cells in fluids. Or cell-free nucleic acid may need to be isolated for viral load testing. Microbes, which are often not as easily lysed as human cells, offer their own unique challenges.

"Ten years ago," Dr. Kaul continues, "we would all have been using hands-on methods that could have taken up to three days to get highly pure, high-molecular-weight nucleic acid for Southern blotting. Now we are mostly focused on PCR or other amplification assays, and we can get by with more fragmented or more crude DNA. So life has definitely gotten easier."

However, Dr. Kaul notes, "Preparation of DNA still takes a lot of technologist time and is fairly work-intensive. So if there were an alternative, that would be great. Which is why a variety of kits are now used by most laboratories."

"Kits for extracting nucleic acids from a variety of types of tissues have proliferated," agrees Mark Sobel, MD, PhD, chief of the molecular pathology section at the National Cancer Institute. "This really started a few years ago with companies like Gentra, who marketed a quick extraction procedure for isolating DNA from blood and some buccal swab preparations. It has expanded to where some companies even make it possible from an experimental perspective to extract DNA from samples with lots of connective tissue components. So I think industry is really helping out in that regard."

Yet kits offer limited help. "While PCR itself is becoming more automated, preparation of samples for PCR is still mostly done by manual methods," says Margaret Gulley, MD, associate professor of pathology at the University of Texas Health Science Center at San Antonio. "Some kits work well and they are helpful, but it is still labor-intensive and expensive and takes a couple of hours."

But Dr. Gulley sees potential delivery at hand. "We are seeing instrumentation coming on board that will help us with that pre-amplification step," she says. "At this time, however, we are right in the middle of a rapid development phase in instrumentation. So it is hard to know when to buy these systems."

Automated instruments that make RNA or DNA have been on the market for about a decade. They have not been adopted more widely because they are not flexible enough, Dr. Kaul points out. "They are well suited for a laboratory doing a high volume of a single sample type, like a genetic testing facility looking only at blood samples and only at DNA," she says. "But in many small clinical laboratories we get all sorts of different sample types and lots of different analytes. These systems may not be able to accommodate that variety." So most laboratories have not purchased them.

"Now," Dr. Kaul says, "commercial folks are catching on to the fact that available instruments don’t suit our laboratories. We don’t want to make 96 samples of the same type at a time. So some flexibility has been built into the newer instruments."

Still, the decision whether to purchase one of the new automated extraction instruments demands hard thought. Says Jeffrey Kant, MD, PhD, director of the division of molecular diagnostics at the University of Pittsburgh Medical Center, "We are not quite at the point of looking at automated ways of handling nucleic acid extraction. We need a certain volume to make that worthwhile and cost-effective." Dr. Kant says he was "pretty impressed" with the schematic of one instrument. "It looked pretty neat," he says. "But it also looked pretty expensive." He sees current instruments as more suitable for commercial laboratories and large pharmaceutical research studies.

Daniel H. Farkas, PhD, HCLD, a former molecular diagnostics laboratory director at William Beaumont Hospital, Royal Oak, Mich., who is now director of clinical diagnostics at Motorola Clinical Micro Sensors in Pasadena, Calif., also considers current automated instruments "incredibly expensive." Affordable automation is sorely needed in molecular diagnostics, Dr. Farkas believes. "After growing rapidly in the ’90s, this field now seems to have reached a plateau," he notes. The reason: "We haven’t experienced the routinization of molecular diagnostics that occurred in clinical chemistry," Dr. Farkas postulates. "There is no black box or instrument where you can put an aliquot of blood or urine in one end and get an answer out the other end."

Many companies are working to devise instruments that will meet this need, among them Motorola Clinical Micro Sensors, where Dr. Farkas heads bioelectronic detection platform development. "We are working with potential partners to integrate technologies that will make molecular diagnostics so routine that it can be done in any part of the clinical laboratory, and ultimately in a handheld point-of-care device," he says.

In the meantime, Dr. Gulley sees a more basic need. "I feel that we need much more education of pathologists about molecular diagnostics," she says. "We could do a lot more to bring molecular testing to the forefront of the pathologist’s attention."

Better cooperation between surgical and molecular pathologists would facilitate molecular testing. For example, Dr. Kaul says, "One sample type we see a lot is paraffin blocks. We can usually get amplifiable nucleic acid out of formalin-fixed and paraffin-embedded tissues, but it can be a real challenge." Improved communication regarding fixation or retention of unfixed tissue would mitigate this problem.

With regard to nucleic acid preparation kits, two companies share the market now, Gentra Systems and Qiagen. "Far and away Qiagen and Gentra Systems dominate the market because they have outstanding products," Dr. Farkas says. Qiagen and Gentra have adopted different basic technologies; each company offers hundreds of kits tailored to preparing DNA or RNA from different specimen types for different applications.

Gentra offers a line of liquid-phase purification kits called Puregene, first introduced in 1992, which purify DNA through a series of salting-out steps using buffers of varying ionic strength followed by centrifugation. Gentra’s Generation product line employs solid-phase chemistry that captures genomic DNA from whole blood or cultured cells added directly to a column. Non-DNA materials are washed off, and DNA is eluted. An important distinction is that Gentra’s solid-phase support is not a silica product, but an adsorbent developed in-house by Dr. Heath.

A liquid-phase product for preparing total RNA, called Purescript, has been marketed since 1995. Solid-phase chemistry for preparing RNA is in development.

Gentra product manager Darin O’Brien, PhD, notes that liquid-phase technology is easy to scale for small to large samples. "Solid-phase chemistries are not as flexible, and you can easily overload these columns," he says. "Their advantage is very high throughput."

Gentra has documented that its kits provide pure DNA (A260/A280, 1.8-2.0) with an expected yield of 35 µg per mL of human whole blood. More than 95 percent of purified DNA exceeds 50 kilobases (kb) in length.

The company also is conducting a long-term study of the stability of DNA purified from whole blood with a Puregene kit and stored at 4°C. Each year an aliquot is removed and tested on a quality control panel (Southern blot, PCR amplification, molecular weight, restriction enzyme digestion). "We know that the DNA performs as well today as it did eight years ago," says Dr. O’Brien. Adds Dr. Heath, "The reason the DNA is holding up so well is that residual impurities that can cause instability have been removed."

At Qiagen, says Tim Fleming, marketing manager for molecular diagnostics, primary product lines include QIAamp, for preparing genomic DNA or viral nucleic acid, and RNeasy, for obtaining total RNA from blood or tissue.

All QIAamp and RNeasy kits employ a silica gel membrane in a tube to which nucleic acid binds under specified salt conditions. Undesired substances are washed off with various ionic strength buffers, and nucleic acid is eluted with low- or no-salt solutions. Nucleic acid from the column is immediately ready for downstream applications without time-consuming pelleting, ethanol precipitation, or rehydration, Fleming says.

Typical numbers for DNA yield are 4-6 µg from 200 µL of whole blood using Qiagen’s Mini product and almost 400 µg from 10 mL of whole blood using its Maxi product. As few as six copies of viral RNA have been detected using RNA isolated with QIAamp kits. "But in the long run it will be reproducibility and consistency that will be important in molecular diagnostics," Fleming says. "That is what we are seeing in all our data."

Qiagen has agreements with Visible Genetics and Abbott that allow them to use QIAamp technology as the front-end extraction and preparation method in their diagnostic systems. "We are aiming toward becoming the industry standard," Fleming says.

Both companies’ kits offer the advantage of nonorganic methods, doing away with the hazard and special disposal procedures associated with phenol-chloroform extraction.

Dr. Kaul has worked with both companies’ products. "Both give pretty high-quality, clean DNA," she says. And the kits provide a quality-controlled product and save labor and time. "We can get DNA from some samples in about one hour in our laboratory," she says, "although typical times are more like two to three hours." But, she notes, "They are still not the final answer."

Dr. Gulley uses Gentra’s Puregene kits to extract DNA from blood and bone marrow, and Qiagen’s silica spin columns for plasma and cerebrospinal fluid samples. "They both work great, but both methods are still very labor-intensive," she says. "Both kits require centrifugation steps and hands-on manipulations, which add up to hours of work."

Putting on his former clinical laboratory director’s hat, Dr. Farkas says, "These kits could improve in price. A significant fraction of the cost of doing the test was the cost of the nucleic acid extraction kits. But they were consistent and gave good results."

Dr. Sobel agrees that both companies’ kits produce good nucleic acid. In addition, he says, "When you use a commercial kit, you can guarantee GMP. You can certainly make up your own mixture and certify it. But then you need to make enough and inventory it to meet standards for laboratory inspection." Dr. Sobel notes that an old-timer like him, who had to make his own restriction enzymes when he started, might tend to do more in-house reagent preparation. "But anybody under the age of 40," he says, "is going to prefer using kits, which are more efficient and provide better quality control."

Kits also have enhanced analysis of tissue in cancer pathology. "Before kits, no one would consider extracting minute amounts of solid tumors," Dr. Sobel says. "What we are doing is streamlining conditions and allowing people to use microscopic amounts of material that we never could have done 10 years ago."

In addition to their other advantages, Dr. Kant says, "The strongest endorsement for these kits is that our technologists tell us that they like them."

Contemporary automated nucleic acid preparation instruments are offered by Roche, Qiagen, and Gentra.

Sharon Sheridan, marketing manager at Roche Molecular Biochemicals, says the company’s Magna Pure LC isolates DNA, RNA, mRNA, or total (for example, viral) nucleic acid. It uses glass beads coated with a magnetic layer. Held inside reaction tips, the beads, with bound nucleic acids, are carried forward by magnets through a series of processing steps. "This is truly a fully automated process," Sheridan says. "Once you load the sample, there is no intervention and no opportunity for samples to get mixed up."

After lysis and proteinase K digestion, nucleic acids are bound to the beads. (For RNA and mRNA, DNase digestion is first done.) Contaminants are removed through a series of wash buffers, and the beads are moved into a heated block in a low-salt environment, where nucleic acids are released. Magna Pure LC can pipette into various reaction vessels for downstream reactions, and it can assemble reagent mixes. (For information on how Magna Pure LC works, see biochem.roche.com/magnapure.)

For mRNA isolation, streptavidin-coated magnetic particles are used rather than glass beads. Reaction buffers contain biotinylated oligo-dT. Biotin binds to streptavidin on the particles, and the polyT of its oligo-dT captures the polyA tail of mRNA. All other nucleic acids are then washed away.

Typical Magna Pure LC DNA yields are 7-10 µg from 200 µL of whole blood and 11-12 µg from 106 HeLa cells. About 77 ng of mRNA can be obtained from 1.2 mL of peripheral blood mononuclear cells. Magna Pure LC produces DNA fragments of about 65kb from whole blood. It can process from 10 to 300 µL of blood.

Sheridan emphasizes there is no cross-contamination on Magna Pure LC for either nucleic acid isolation or PCR reaction setup. For reproducibility, she quotes a coefficient of variation of less than three percent across 15 pairs of instruments (Magna Pure LC plus Light Cycler). Cost of Magna Pure LC is $85,000; per-sample cost (reagents plus disposables) is about $2 for DNA, $3 for RNA, and $4 for mRNA.

Magna Pure LC takes 90 minutes to purify 32 samples and will process from one to 32 samples. One customer, who had been running four batches per day with manual kits, saved 0.5FTE per day after switching to a Magna Pure LC, Sheridan says.

Qiagen’s nucleic acid isolation technologies are automated in the BioRobot 9604 system, which was initially developed for automated RNA extraction for viral load assays. Current models can also extract genomic DNA from blood and buccal swabs; they can do amplification setup as well. As with QIAamp kits, no final rehydration step is needed. Throughput is 96 samples in about two hours.

BioRobot 9604 uses a centrifugation step for final nucleic acid elution. "You manually transfer a 96-tube plate into a centrifuge," Fleming says. "We would like to move out of gravity and toward vacuum. Ideally," he adds, "we would also like to wed the BioRobot 9604 to an existing automated amplification instrument."

List price for BioRobot 9604 is just over $100,000. Fleming cites a per-sample cost (disposables plus reagents) of "a few dollars per sample." He acknowledges that "BioRobot 9604 is still more appropriate for large reference laboratories." Fleming claims that some large laboratories have seen a reduction of three to four FTEs after adopting BioRobot 9604, though he didn’t name them.

The newest kid on the block is Gentra’s Autopure LS, which debuted at a trade show in October. It incorporates Puregene liquid-phase chemistry for purification of genomic DNA from whole blood. Autopure LS, which processes 96 samples in eight hours, is a true walkaway instrument, according to Dr. O’Brien: "You transfer blood to the input tube, and about one hour later you have purified DNA."

Dr. O’Brien says the Autopure LS’ main advantage is its ability to handle larger volumes than other instruments, 1-10mL, which is ideal for customers who want a lot of DNA that can be archived-contract research organizations, pharmaceutical companies, and biorepositories. But one of the first customers was a clinical molecular diagnostics laboratory that also supports a substantial amount of clinical research. U.S. list price for Autopure LS is $195,000.

Few laboratorians consulted could say anything about automated nucleic acid preparation instruments beyond the near-universal comment, "I’d love to have one, but I can’t justify the expense." Dr. Gulley’s comments are typical: "They are outrageously expensive. This is a problem, because most molecular laboratories can’t afford these instruments unless they are reference laboratories bringing in high volumes."

Directors at two reference laboratories offered limited evaluations.

Elaine Lyon, PhD, medical director of molecular genetics at ARUP in Salt Lake City, says her group began last August to conduct a validation study with Roche on the Magna Pure LC, and it purchased the Magna Pure in December. "It is just now getting into the clinical laboratory," she says. Magna Pure LC will be used to extract DNA from peripheral blood leukocytes for genetic analysis for factor V Leiden, hemochromatosis, factor II (prothrombin), and methylene tetrahydrofolate reductase, or MTHFR.

For ARUP, Magna Pure LC’s main attraction is that it extracts DNA and loads it into the Light Cycler with only one intervening centrifugation of the Light Cycler carousel. The centrifugation step doesn’t require technologists to handle individual samples; the entire carousel with 32 samples is centrifuged. A single bar code can be used from specimen accession to result. "Whenever there is manual manipulation, there is concern that a sample can get switched," Dr. Lyon says. "We wanted to eliminate that concern."

Validation studies compared Magna Pure LC to the Qiagen BioRobot 9604 that ARUP had been using. Magna Pure LC "matched or exceeded" BioRobot 9604 in DNA yield and purity, Dr. Lyon said. Rate of dropouts-samples that would not amplify-was similar with both instruments, one to two percent.

ARUP may use its BioRobot 9604 for non-Light Cycler applications, such as beta-globin sequencing for hemoglobinopathies and cystic fibrosis.

At Specialty Laboratories in Santa Monica, Calif., a Qiagen BioRobot 9604 is used to extract HIV and HCV RNA from plasma for viral load and sequencing, says Andreas Bakker, PhD, director of genotyping and molecular technology. Qiagen BioRobot 9604 requires a full platform-96 samples-for each run, Dr. Bakker notes, so it is cost-effective only for large-volume laboratories. "Our DNA assays have also reached a volume that justifies automation," Dr. Bakker says, "so we are currently evaluating instruments for that application as well."

Dr. Gulley remains hopeful that automated extraction instruments will evolve to be cheaper, faster, and better, as they have in other areas of chemistry, so that more laboratories can afford them. "Just as it happened for protein, it will happen for RNA and DNA," she predicts optimistically.

To extend the reach of kits and instruments, communication between molecular and surgical pathologists needs to be improved. "A major type of sample we are often asked to look at is archival tissue in paraffin," Dr.Kant says. "There are adaptations of kits that seem to work satisfactorily. And we have a homebrew method that works pretty well." But it is important for pathologists to appreciate that the best type of specimen for molecular analysis is fresh or frozen tissue, not paraffin-embedded specimens after formalin fixation. "You are at a disadvantage having to work with nucleic acid out of paraffin," Dr. Kant says. "I know the Europeans have made a real effort to make sure that specimens are obtained correctly."

But the answer is not simple. Dr.Kaul notes that alcohol-based fixatives give up beautiful nucleic acid but do a poor job of preserving histology. Conversely, she says, "Hematopathologists love B5 fixative, but you can never get DNA out of that."

Another complication is that it is not always obvious ahead of time when a molecular analysis will be needed. "Molecular pathology is a tertiary technology for most cancer applications," Dr. Kant says. It has a role to play now in selected cases where sections and immunostaining don’t provide an answer.

Dr. Kaul notes other problematic tissues. Bone presents problems for nucleic acid extraction because it is difficult to get the cells out of the trabecular structure. "In order to prepare slides, bone needs to be decalcified using a lengthy soak in an acid solution so that the tissue can be cut. But this acid soak destroys the DNA," she says. Lymph nodes are often put into B5, but then they are suboptimal for gene rearrangement studies or for PCR to look for a microbe (e.g. M. tuberculosis) if granulomas were present, for example.

But minor problems like fixatives disappear when molecular pathologists envision an ideal nucleic acid processing instrument. "What we really need," Dr. Gulley says, "is for these instruments to be able to handle any sample type-blood, plasma, cerebrospinal fluid, urine, pleural or other bodily fluids, bone marrow, and biopsy tissue. In the future," she adds, "I imagine that one instrument would do extraction as well as amplification and be linked to a computer for bar-coding and reporting."

Dr. Sobel foresees that both macro instruments and microarrays will perform these functions. "That will be incredible," he says. "That will be like what we considered science fiction in the past. It won’t happen in the next year or two, but in the next decade it will be done."

William Check is a freelance medical writer in Wilmette, Ill.