Back to Basics:
In this 2017 “Primer” series, we are following a somewhat different procedure than we did in the column’s first four years. Every odd-numbered month features a “Back to Basics” article in which topics that were covered in earlier installments are revisited and expanded. In the even-numbered months, new topics are discussed.
So, we began this year’s series, in the January 2017 issue, with the first of the Back to Basics articles: an overarching view of nucleic acid amplification technologies as applied in the clinical laboratory. As I pointed out then, there are a wide range of uses for this technology in current molecular diagnostics (MDx) methods. For the non-MDx specialist, there is one particular type of polymerase chain reaction (PCR) application that is probably the most frequently encountered.
As such, in this March 2017 issue—an odd-numbered month, the third month of the year—it will be worthwhile to go back to the basics to review this type of MDx approach in a bit more detail than the January overview provided. I hope that an appreciation for the method and its intrinsic strengths, weaknesses, and interpretational nuances can help guide readers to have more meaningful communications with their respective MDx lab service providers in this context.
For this discussion, I will select this most frequently encountered PCR application as real-time, fluorescent, probe-based qPCR detection of absence or presence/quantity of a single target. I will choose an infectious disease agent as the target example. Of course, this example will not be the single most common form of MDx assay encountered for all readers; however, in terms of the number of types of assay on the market, assays of this type almost certainly underlie the majority of lab reports for the majority of readers. I will also show, later in this article, how the critical concepts that apply to this instructive example can apply with very little additional concepts to a number of method variations. Together these will cast a very wide net.
Performing the test
Our basic example method really starts with a hypothesis on the behalf of the attending clinician. Based on the patient’s presentation, a particular pathogen or perhaps type of pathogen is suspected as the etiologic agent. Based on this suspicion and by reviewing the available laboratory test menu, the collection of an appropriate sample type can be performed. The importance of this step should not be underestimated, and if you’re the requesting clinician and you’re unsure, it’s a good idea to contact the lab and verify appropriate sampling method, collection media type, and other details before collecting and sending a specimen. In my experience (and supported by discussions with other laboratorians), incorrect sample types arrive in the MDx lab with unfortunate regularity, and must be rejected. While the difference between “tube top color,” for example, may not seem overly significant to the sample collector, it may have major impact on the underlying assay chemistry to be performed. Even when a variation in sample type isn’t explicitly known to be problematic, if the sample as submitted does not match a sample type which has been appropriately validated for the assay, the assay can’t be meaningfully performed.
Now we shift to the MDx lab, which has in hand both a requisition for a particular test and an appropriate matched specimen type. The lab workflow proceeds through a nucleic acid extraction, which purifies total DNA and RNA (derived from host/patient, possible pathogenic agent, as well as any contaminating organisms) and ideally leaves behind all manner of other and potentially inhibitory substances. In a simplified view, this extracted nucleic acid is then mixed with two PCR primers, a fluorescent probe, and a buffer containing a DNA polymerase. The primers are short synthetic single-stranded DNA oligonucleotides which share sequence identity with the target of interest; they’re spaced a few hundred base pairs apart on the target, one on each target strand, and they “face” each other in a DNA polarity context. The probe is a similar short single-stranded DNA molecule, with attached fluorophore(s), and it shares sequence identity with a region of the target found between the two primers.
There are different types of probes, but all work by exhibiting a large change in fluorescence intensity when they anneal to their complementary sequence as found between the two primers. This mixture is then subject to a thermal cycling process which allows primers to bind their cognate target (if present) polymerase to create a target copy, and probe to bind the resulting target copy. The method is exponential over cycles, meaning that even a tiny number of correct target molecules (if present) will lead to a very large number of replicated “amplicons” and subsequent bound probes. This process is done in a translucent or transparent vessel, under the watchful eye of a fluorescence measuring system. If the target is not present in the reaction, no increase in fluorescence is observed; if the target is present, an exponential (initially, before exhausting reagents and reaching plateau) increase in fluorescence is observed. Thus, an increase in fluorescence corresponds to a positive result for the target of interest; it’s that simple.
If there is more target material in the sample, this increase in fluorescence occurs earlier (after fewer thermocycles) than if there is less target material. It’s a log-linear relationship, and if quantitative results are desired, it’s necessary to run a parallel set of known target positive reactions at defined concentrations; these create a standard curve of fluorescent signal vs. cycles against which the patient sample can be interpreted. For a purely qualitative assay, we can dispense with a standard curve (although it’s interesting to reflect that the method is intrinsically qualitative, and that in this case we’re just not extracting this information).
Analyzing the result
The lab now can result back a response to the initial requisition for a test on the target of interest—absence/negative, or presence/positive (and possibly quantity). What are some of the issues the ordering clinician should bear in mind when looking at this report? A false negative can arise from several causes. These could include wrong sample type (hopefully already identified and rejected by the lab, but not always); or genetic variation in the target organism (a significant sequence change under either of the primers, or under the probe, will cause the assay to fail); or perhaps the target organism is present but at low numbers, below the LOD (limit of detection) for the assay. Inhibitors in the sample could also cause a false negative. (However, the assay will generally include an internal control (IC) in the form of another, definitely present genetic marker. Detection of that marker serves as an assurance that assay inhibition has not occurred, while conversely lack of IC detection will flag an inhibited, invalid assay.)
False positive results can also occur, although one of the reasons for probe-based PCR’s popularity is that by requiring three perfect (or near perfect) sequence matches to the target—one each for the primers, and one for the probe—cross reactions or poor specificity are minimized. While unexpected reasons for fluorescence increase during the assay can exist, such as nonspecific nuclease activity degrading probes and releasing active fluorophore for detection, the kinetics of such processes generally looks quite different from an actual true positive with its exponential and plateau (sigmoidal) curve. In the case of very weak (late cycle) signals, however, this can be more difficult to distinguish.
Further applications
As noted above, this model could be expanded upon to cover a wider range of assays in use. It could, for example, be applied to genomic targets, such as particular genetic markers (present or absent). No particular changes in our process occur in this case, although generally this means that LOD issues are not a concern, as a haploid marker will occur at easily detectable numbers. (Applications such as detection of microchimerism, or minimal residual disease in oncology, do however need to consider LOD.)
Another possible application: looking for an RNA target, either viral or an endogenous mRNA. All of the above methods and issues apply; the only changes are in the underlying laboratory work. Sample extraction and handling under special precautions to preserve RNA integrity, and the addition of a reverse transcriptase enzyme which converts the RNA to a DNA copy which is then detected exactly as above, are required; also, the IC should be RNA to reflect the desired sample type.
Another application is multiplexing, that is, the ability to look for several unrelated targets in the same reaction at the same time. This is readily handled “behind the scenes” by mixing the unrelated primer and probe sets in one reaction, as long as the probe fluorophores are at distinguishably different wavelengths so that the real-time PCR instrument observes them independently. In fact, most “single target” real-time PCR assays are multiplex, with the IC already being run as a second multiplexed signal.
Multiplexing by this approach is, however, limited by this fluorophore differentiability issue, with most instruments being capable of resolving six or fewer simultaneous channels or reactions.
Whether you’re testing for an infectious disease or a specific genetic marker, whether the target is DNA or RNA in nature, if the MDx test method you’re requesting is looking at a limited number of targets, it’s likely that some variant on probe-based, real-time PCR will be performed by the lab. While these methods have truly impressive specificity, sensitivity, and flexibility, no test method is infallible. An understanding of the overall method workflow and some of the possible causes of false positive or negative results is thus beneficial in appropriately interpreting test results in the full clinical context.
