Skip to main content
Premium Trial:

Request an Annual Quote

Cell-Free DNA Purification Kit Comparison Shows Spin Column Tech Superior


NEW YORK (GenomeWeb) – A group of researchers led by the Macquarie University in Sydney, Australia have compared a variety of commercial kits on their ability to extract circulating cell-free DNA (cfDNA) from patient blood samples.

The team found that, in general, spin column-based extraction kits — such as Qiagen's QIAamp circulating nucleic acid kit and Norgen's Biotek's Plasma/Serum cell-free circulating DNA Purification midi kit (NorS) — performed better than more recently introduced magnetic bead-based kits.

To detect cfDNA, researchers normally use tools based on either spin column or magnetic bead technology. While spin columns are more expensive and require much longer to produce results than magnetic bead-based methods, they usually produce higher recovery rates of cfDNA in patient samples

In a study published in Cancer Genetics last month, the researchers examined the ability of several manual and automated spin column- and magnetic bead-based commercial kits to detect spiked-in melanoma tumor cfDNA in plasma from healthy patients.  

"Our initial goal was to increase recovery of circulating tumor DNA (ctDNA), playing around with a few different parameters, such as incubation time, changing elution volume, or whether to repeat elution or not," Macquarie University senior research fellow and first author Russell Diefenbach explained. "Disappointingly, the [parameters] didn't really change the yield dramatically, as we also wanted to improve the sensitivity or limit of detection of the DNA for downstream activities."

Once Diefenbach and his team showed how little they were to maximize ctDNA yield with the Qiagen (QiaS), they decided to compare the kit to tools developed by other firms. The researchers started by spiking three of four milliliters of healthy human blood samples with 500 nanograms per milliliter of cfDNA purified from BRAF mutant melanoma cell lines.

"Initially, we spiked in a high amount [of cfDNA] because we wanted to capture information across a broader range or spectrum of DNA fragments," Diefenbach said. "We also decided to use healthy patient samples because we didn't want to waste precious patient DNA in this type of analysis."

The team ran blood samples from the Melanoma Institute of Australia  — along with the remaining 1 milliliter of unspiked blood samples — through the commercial cfDNA purification kits. In addition to Qiagen and Norgen's spin column-based kits, the researchers used magnetic bead-based kits including Qiagen's QIAamp minElute ccfDNA mini kit (QiaM), Promega's Maxwell RSCS ccfDNA plasma kit (ProM), Applied Biosystem's MagMax cfDNA isolation kit (ABioM), and Bioo Scientific's NextPrep-Mag-cfDNA isolation kit (BSciM).

"There seems to be a new kit almost every week to extract ctDNA," Diefenbach said. His team therefore aimed to select the best representatives of magnetic and spin column-based kits available on the market. He noted that "this [study] is not meant to capture all kits, as there might be kits out there that people are developing, but to standardize the field if possible."  

For spiked-in plasma cfDNA experiments, the team first extracted cfDNA from the melanoma cell line M229 using the QiaS kit.

The researchers then analyzed purified spiked-in ctDNA with lengths ranging from 50 base pairs (bp) to 800 bp. While Diefenbach estimated that the length of normal ctDNA to be about 130-170 bp, he also noted that multiple copies of ctDNA can bind to each other and form longer bp. The group then used a droplet digital PCR system to detect non-spiked wild-type NRAS or spiked-in mutant BRAF.

Afterward, the team used lower amounts of spiked-in ctDNA for cell types that were closer to physiological levels normally found in patients, Diefenbach said.  

The team saw that all purification kits recovered varying amounts of low molecular weight DNA fragments, ranging from 50 to 766 bp. While the researchers failed to collect 25-bp DNA fragments because of protein "digestion" in the bloodstream, Diefenbach noted that his team managed to recover the 50-bp fragments, "albeit at lower efficiency levels."

The researchers found that the QiaS, NorS, QiaM, ProM, and ABioM kits had consistent levels of recovery across the range of 75-808 bp. In contrast, they saw that the BSciM kit displayed the most variable profile of recovery.

Meanwhile, the researchers noted that the spin column-based QiaS and NorS kits demonstrated no major difference in overall recovery based on mean recovery of total DNA fragments. In addition, they saw that both tools showed a "significantly greater recovery rate" than the magnetic bead-based kits.

In terms of the magnetic-based kits, the researchers saw that the QiaM performed significantly better than its counterparts at extracting tumor DNA fragments and minimized contamination of cells and other small molecules.

Diefenbach and his colleagues then compared the QiaS kit with the QiaM kit, based on their total recovery efficiency using spiked-in DNA. The group analyzed the tools based on extraction of either non-spiked cfDNA or spiked-in cfDNA carrying the BRAF somatic mutation from healthy plasma. They found that the average DNA copy number extracted from the plasma of five healthy patients dropped significantly in the QiaM kit compared to the QiaS tool.

Overall, the study authors argued that currently available cfDNA extraction kits — regardless of the type of technology used — did not "surpass" the QiaS kit in terms of improving the sensitivity of cfDNA detection.

Diefenbach explained that while the QiaS is a little more expensive and time consuming, the tool's availability in an automated form for high throughput use — known as the QIAcube — helps distinguish it from the other spin-column based kit in the study.

"I'm not saying that we shouldn't use the NorS system, but [rather] it depends on your downstream applications," Diefenbach said. "If you're interested in high throughput, the QIAcube is a better set up, versus a manual kit like the NorS."  

However, the team also noted that if researchers used alternative, cheaper, and less time-consuming methods like a magnetic bead-based kit, such as the QiaM kit, they would only have to accept that the overall yield of recovered ctDNA may be less than expected.  

To make the ProM kit a potential alternative, the study authors argued that the larger sample, volume-based option would need to perform well as the QiaS in collecting non-spiked DNA and only "be available in an automated mode."

Moving from comparison studies, the researchers will use the QiaS extraction tool for downstream next-generation sequencing technology because of the assay's ability to multiplex a variety of cancer mutations. Diefenbach said his team is interested in developing "customized NGS and targeted hotspot panels" that will identify certain cancer biomarkers —as well as understanding why patients are resistant to immunotherapies — in future functional studies.

According to Diefenbach, researchers might also be able to apply the extraction tools to purify other species of DNA circulating in the plasma.

Markus Sprenger-Haussels, director and head of sample technologies at Qiagen, said that the firm's QIAamp kit — which costs A$36.70 ($26.04) per two milliliters of sample — distinguishes itself from the other company's tools because of multiple "little differences" the firm's developed over several years. These changes include "well-balanced lysis chemistry to release circulating cfDNA from its nucleosome complexes, as well as from extracellular vesicles, exosomes, and oncosomes."

"If you want to extract circulating cfDNA from plasma, it's best to use your plasma and real circulating cfDNA, rather than using artificial test systems," Sprenger-Haussels said. "Other companies start by making material like an artificial plasma, which has a similar protein composition, but in reality, it acts differently. You have a lot of background material that might interfere with sample prep, which will not be optimized to deal with [the sample]".

Sprenger-Haussels noted that the Qiagen is now working with customers that want to use 10 milliliters or more per extraction, which requires more time than the standard QIAamp tool. The firm therefore developed the QIAamp minElute magnetic bead tool, which costs A$18.66 per run and is  more scalable for larger volumes of blood.

According to Sprenger-Haussels, different researchers have argued as to which DNA bp length teams should focus on to detect the presence of tumors in a patient's bloodstream. Some claim that that tumor cfDNA is very small, and if there are larger concentrations of DNA in the plasma, it stems from the tumor rather than DNA, Sprenger-Haussels said. However, other researchers claim that the DNA of interest for cancer is in the higher molecular weight fraction.

"Therefore, Qiagen has tried to optimize protocol for both groups [with its tools]: one group focusing on small DNA, and other teams seeking large DNA," Sprenger-Haussels said. "We do not exclude large DNA, as we have learned that in this fraction valuable tumor DNA might be present, originating from oncosomes and the degradation process of circulating tumor cells."

Promega Senior Research Scientist Douglas Horejsh found it peculiar that the researchers initially spiked an abnormally high amount of naked DNA than would be typically found in a patient’s plasma. He explained that endogenous levels of cfDNA are usually around five to 25 nanograms per milliliter, rather than the overloaded 500 nanograms per milliliter in the spiked samples performed in the first portion of the study.  

"When we normally work with researchers or customers, we typically urge them to compare performances between platforms using matched plasma with no spikes," Horejsh said. He argues that using spiked systems can lead to "poorly defined" results and different interpretations, based on what researchers are trying to demonstrate with the tools.

The firm's automated ProM kit requires the user to pipette 0.2 to 1.0 milliliters of plasma into a cartridge before inserting it into a machine. Costing $31.26, the extraction tool then runs for about an hour to effectively purify the plasma sample.

However, Horejsh acknowledged that Promega is working on the Maxwell kit to improve volume size for extraction. In addition, the firm plans to release kits to increase volume sizes on "other, larger automation platforms" in the future.

"Typically, researchers would like to use one tube of whole blood, which yields four to five milliliters of plasma," Horejsh said. "We are working to fill this need in a simple-to-use format."

Norgen Biotek VP of business development Nezar Rghei does not understand the researchers’ rationale behind their decision to choose the QiaS as the golden standard for ct-DNA extraction. .

According to Rghei, Norgen's spin-column kit differs from Qiagen's because it employs a different type of matrix. Unlike Qiagen's silica-based column, the NorS uses silicon carbide in order to bind tumor nucleic acids.Like Horejsh, Rghei believes that the way the researchers spiked the DNA, "naked" and without histones, as well as the amount of tumor DNA does not accurately reflect the amount present in a normal patient's blood sample.  

"It would make more sense to push the concentration of spiked DNA at normal levels, to observe more differences on the Qiagen tool," Rghei said. "When you spike the sample, the DNA is degraded very quickly by enzymes … and [smaller] bands may have been digested because they were not covered or protected by proteins against nucleases."

According to Rghei, Norgen'ssample collection kit requires about 80 minutes to produce results and costs about $33.

Overall, Diefenbach highlighted that the study is the first of its kind to use a wide size distribution of spiked-in DNA fragments to examine commercial cfDNA kits.  

The study authors added that "establishing the performance of commercial cfDNA kits for recovery of variable sizes of DNA is important for optimizing the utility of circulating DNA as a cancer biomarker."