Quantifying circulating microRNAs in biomarker research can be a sticky business, according to a group of researchers who highlighted in a report last week a host of sample-collection and -preparation factors that can influence the accuracy of quantitative PCR miRNA studies.
In a study published Dec. 7 in the Journal of Molecular Diagnostics, the team described a variety of confounding factors, such as the type of collection tube used in sampling, the influence of added anticoagulants and stabilizers, sample size, and the role of endogenous serum factors that co-purify with RNA and inhibit polymerases.
The researchers also made specific recommendations on how to better optimize detection. Putting them into practice, they were able to improve the sensitivity of miRNA detection in their tests up to 30-fold in some cases.
Dominik Duelli, a molecular pharmacologist at the Chicago Medical School at Rosalind Franklin University and lead investigator of the study, told PCR Insider this week that his group became interested in optimizing and standardizing quantitative RT-PCR miRNA detection after beginning research into circulating miRNAs as potential biomarkers in breast cancer.
Several companies, including Life Technologies and Exiqon, now offer products for detecting miRNA in blood samples for biomarker development and disease research. With qPCR currently the choice approach, the field needs assurance that measured concentrations actually represent the amount of miRNA in a sample, and "such assurance is often lacking," the authors wrote in their paper.
Using qPCR-based techniques and considering the factors highlighted in the paper could enable quantitation of low-abundance miRNAs when usual techniques fail, the researchers wrote, and could "increase the repertoire of miRNAs that can be analyzed as potential biomarkers of disease."
Duelli said the group initially wondered whether the collection tubes they used for their work might influence miRNA concentrations and their ability to amplify and detect the molecules.
"That led us into looking at a whole bunch of other things: Maybe it doesn’t matter just how you collect, but how much you collect and how much you measure … and what can you do in cases where you don't have a choice of what tube to use?," Duelli said.
"That's how we got into it, just starting to consider all sorts of things."
In investigating sample collection and endogenous co-purifying factors, and attempting to optimize their process, the group quantified two miRNAs, miR-16 and miR-223 — which are common targets in cancer and other disease studies — in blood samples collected from a number of sources, including healthy adults, samples from the Susan G. Komen Foundation for the Cure tissue bank, and samples supplied by the AIDS Cancer and Specimen Resource at the University of California, San Francisco.
The researchers tested a variety of sample-collection tubes — some without anticoagulants and others containing anticoagulants like EDTA, heparin, sodium citrate, or NaF/KOx. They also evaluated whether separately adding anticoagulants and blood stabilizers could offer additional benefit.
They reported the best results using tubes containing NaF/KOx. Though both miRNAs have very different abundance in blood, recovery and accurate detection for both of them depended on the blood-collection method, the team wrote, suggesting that collection procedures should influence additional miRNAs regardless of their abundance.
Tubes containing heparin, which the group used when beginning its research, are a bad choice, Duelli said, because heparin binds RNA and can interfere with polymerases.
But, he said, if researchers are limited to heparin tubes, his group found that adding the heparin-digesting enzyme heparinase to the samples can increase the ability to detect miRNA.
Additionally, adding NaF and KOx later in the process can also benefit detection. Added together, the anticoagulants increased miR-16 detection twofold in plasma collected using EDTA tubes, and threefold in serum, the researchers wrote.
With the influence of sample collection methods in hand, Duelli's group then went on to investigate issues in the composition of blood samples themselves that can interfere with qPCR.
The group applied a variety of approaches to improve the purity of isolated RNA and increase the sensitivity of miRNA detection.
The first approach, Duelli said, was to "just purify the plasma really well." Adding "a single acidic phenol/chloroform extraction step, followed by adsorption of RNA on silica membranes reduced the interference by blood-borne polymerase inhibitors and effected a 4.4-fold increase in miRNA quantitation," the group reported.
Duelli said that most researchers currently "just take plasma and do a TRIzolic extraction, which is the most common and easiest way to extract RNA."
"But if, for example, we added a couple more phenol/chloroform steps, that helped," he added.
However, Duelli said that repeated extraction and purification steps can become expensive and time-consuming. Also, "each time you handle a sample … the more variability and losses you get."
As such, the group profiled additional polymerases that might avert interference from blood-borne inhibitors without the need for repeated purification.
"What really inhibits the detection of these microRNAs," Duelli said, "is that Taq polymerase, which is the cheapest and most-used polymerase in PCR, be it quantitative or otherwise, is inhibited by several things that can co-purify with RNA."
He said that the researchers decided to try a Taq polymerase mutant, Hemo KlenTaq, developed by a team at Washington University School of Medicine in Saint Louis, to resist the binding of inhibitory molecules.
"They came up with the idea to figure out where these inhibitors bind to the polymerase, and make a mutant [to get] rid of that fragment," Duelli explained.
"But it turns out that [fragment] is required for good proofreading. So while you didn't have to purify your sample so much anymore because now the inhibitor is no longer bound to the polymerase, and you get a lot of amplification of your product, the problem is if you cut off the [binding fragment], you no longer have good proofreading."
"That may not be a problem if you just run a gel because of the size of your band. But if you go to quantitative PCR, you may not exactly be able to see the quantitation," Duelli said.
To overcome this issue, Duelli's team took a combinatory approach, adding an intact Taq polymerase to overcome the reduced proofreading activity of Hemo KlenTaq. "We thought, 'Why not combine both worlds, take the polymerase that they developed to be really good for blood, and add a regular polymerase,'" he said.
The combination of polymerases amplified specific PCR products well enough for qPCR using SYBR Green or TaqMan, the group wrote. "You get the huge amplification from the Hemo KlenTaq, but at the same time adding the normal polymerase, which is partially inhibited, can really help in reducing the background amplification of non-specific RNA," Duelli said.
The researchers tested whether their work to overcome endogenous inhibitors would allow higher miRNA quantitation, analyzing circulating miRNAs in six healthy subjects.
"We looked at the overall abundance of these two microRNAs that we mostly look at," Duelli said, "and what we found is, in some cases, it's very linear, very clear: in five out of six people you get up to 30-fold more amplification products, so you increase the sensitivity and quantity quite a bit."
However, he said there was one subject where the increase wasn't equivalent. This suggests that differences in plasma composition from individual to individual can yield different miRNA measurements, the group reported.
"MicroRNAs as diagnostic and prognostic markers have just taken off like crazy and I think for good reason. But I think we have to be very careful that we can validate these things," Duelli said.
"As far as I can tell, [when] you go from one lab to another things are not as standardized as proteins and mRNA, and that's because the field is still relatively young with a lot of excitement."
Duelli suspects many others have hit on some of the same issues and solutions his group presented in their paper. "I'm guessing other people know similar stuff, but I'm hoping the paper will help people who are just starting; or, for people who have done it for a while, maybe they will think to try some of this."
Overall, he said, the team identified two take-home messages from the results. "First, if you start a study fresh, really look at which particular blood collection tube you use … it really matters what chemicals you have in there because they interfere either with the [reverse transcriptase] or the PCR itself."
"Then it also really depends on how much starting material you use."
Counterintuitively, smaller sample sizes actually produced better results for the researchers. This makes sense, Duelli said, because as you increase your sample, you also increase the endogenous inhibitors that the team showed can reduce miRNA detection.
"So figure out the tube, figure out what the best volume is, and then decide if you want to purify it, which might be costly but very good," he said. "Or do you want to use one of these fancy polymerases?"
Because not all the samples from the six healthy subjects that the group tested showed the same level of improvement with the optimized sampling and preparation protocols, Duelli said, it is likely there are other blood components that the group was not able to account for, perhaps even individual-specific factors like age and habits.
In that case, "what could help is really just to standardize the blood-collection practices," he said. "For epidemiological studies, these things may even themselves out," but they may matter for the individual studies, he added.
In their own research, Duelli said his team will plan to use NaF/KOx tubes, and the combination of Hemo KlenTaq and regular Taq polymerase. In the paper they rate all of the possible steps in terms of time, sensitivity, and cost, to allow other groups to pick and choose a path based on their own needs.
Duelli did not suggest that his study invalidates qPCR technologies currently being used for miRNA biomarker research. Many groups may have come to include the steps his team recommends on their own, he said. However, he said standardization across the field is still necessary.
"Typically what you hear people say these days is, 'microRNA quantitation in the plasma is trivial,' and I agree. It is extremely trivial to get data," Duelli said.
"But for that data to actually reflect what the microRNA concentration is in the blood is absolutely not guaranteed unless you take [some] steps … and [address] the variables we identified."
Have topics you'd like to see covered in PCR Insider? Contact the editor at mashford [at] genomeweb [.] com.