The real-time reverse transcription polymerase chain reaction (qRT-PCR) is often described as the gold standard for measuring differences in gene transcript quantities. However, there are significant problems caused by variability of the RNA template, assay design, data normalization, and inconsistent data analysis. What are some ways to deal with these issues and improve standardization?
Drs. Chittur and Tine
RNA integrity is a major factor in your qRT-PCR experiments. We use both a Bioanalyzer and a NanoDrop to evaluate our RNA quality since both measures offer an independent mode of evaluation. RNA can self-prime itself in an RT reaction as seen in studies conducted by ABRF’s Nucleic Acids Research Group.
In terms of assay design, a combination of oligo-dT and random hexamers works best. With respect to data normalization, most people routinely tend to use GAPDH, b-Actin, or other housekeeping genes (HKG). However, we would caution against this. One should evaluate the HKG to ensure that its expression level is indeed constant across the experimental conditions.
A better choice of normalization would be a set of endogenous genes (including HKG) that do not change in the experiment. The best way to deal with these issues and improve the overall quality of the qRT-PCR data reported in the literature is for investigators to make a commitment to follow the MIQE guidelines.
The goal of these guidelines is to provide a full accounting of the design of the assays employed, the specifics of template preparation and quality control, a detailed description of the experimental conditions for all steps of the assay, specific information on the detection reagents (primers, probes), and details on the analysis of data. Adherence to MIQE guidelines will enforce a more rigorous approach to the application of this technique. It will not only strengthen the research of those who use it but also make it much easier for others to make appropriate use of published assays.
The first effort to standardize molecular diagnostics was recently taken in Europe through the SPIDIA project (www.spidia.eu), which is sponsored by the European Commission. SPIDIA focuses on the pre-analytical steps, which is the main source of confounding variation. For academic applications, the MIQE guidelines indicate the information that should be provided with a report.
Strategies to estimate the confounding variation of the experimental steps have also been described, and the statistical tool is available in software such as GenEx (www.multid.se). This identifies steps where reproducibility is poor. Measures to improve reproducibility can then be taken.
After over 15 years of experience with qPCR, I have come to the conclusion that if the following guidelines are adhered to, one is almost guaranteed high-quality qPCR data. First, RNA of high integrity must be used. I recommend using an Agilent Bioanalyzer on each sample. Typically, RNA with an RNA integrity number of 7 or higher should be used.
Second, accurate determination of the RNA concentration is important so that equivalent amounts of RNA are added to each RT reaction. The NanoDrop spectrophotometer works well for most applications. However, for low levels of RNA, I recommend using a fluorescence-based detection method such as the Qubit® Fluorometer from Invitrogen.
Finally, a suitable reference gene should be used for normalization. Suitable is defined as a reference gene whose expression does not change within the experimental condition (e.g., treated vs. untreated, normal vs. disease). It is acceptable, and often recommended, to use multiple reference genes for normalization.
Of the utmost importance is a high degree of accuracy and precision with pipetting. Conscientious pipetting using automatic (single and multichannel) repeating pipettes are essential to generating high-quality qPCR data.
Inherent to real-time PCR as a method of quantification is the fact that abundant templates (including cDNAs) reach their Ct values and plateau in fewer cycles than rare templates. But as soon as the first amplification product reaches its plateau, the reaction stops for all amplification products because the polymerase binds to the double-stranded DNA product and this results in feedback inhibition, which stops further amplification. Consequently, there is no simple way to measure both an abundant RNA and a rare RNA in the same symmetric reaction unless the primer concentrations for the abundant cDNA are reduced to prevent generation of too much double-stranded DNA.
Needless to say, these problems become even more complex in multiplexed symmetric RT-PCR reactions. The good news is, however, that all of these problems are overcome using RT-LATE-PCR because the amounts of double-stranded DNA generated for each amplicon are low and the polymerase does not bind to the abundant single-stranded DNA generated for each amplicon.
A second major source of difficulty encountered with all one-step RT-PCR methods comes from the fact that reverse transcriptase also exhibits polymerase activity. Unlike the polymerase activity of the Taq, which is suppressed during the RT step by a hot-start antibody, the polymerase activity of the reverse transcriptase is active during the RT step and can cause mispriming. This problem is exacerbated by the relatively long time of the RT step, the relatively high concentration of the enzyme, and the high abundance of the primers used for both reverse transcription and subsequent amplification.
The problem of mispriming during RT is made more difficult by multiplexing. RT-LATE-PCR alleviates but does not entirely eliminate this problem by utilizing a limiting primer whose concentration is significantly lower than the excess primer.
The third major problem in RT-PCR comes from the nature of RNA. It is more easily degraded than DNA and it exhibits a high-level secondary structure. The problem of degradation can be minimized by scrupulous attention to cleanliness and storage. In addition, we have found it advantageous for small samples (down to single cells) to carry out sample preparation, reverse transcription, and PCR amplification in a single tube via a series of dilution steps in order to minimize the risk of sample loss and sample contamination.
The problem of secondary structure is sequence dependent and has to be understood in detail because hairpin structures in the RNA decrease the effective Tm (melting temperatures) of primers to an RNA template as compared to primers to a DNA template.
RT-PCR and long-range PCR reactions have been designed to detect and distinguish inversions, deletions, and duplications. What are the advantages of this approach and how are you applying it to your research?
I have had a fair amount of experience with long-range PCR, not so much for mutation detection but for obtaining large segments of genomic DNA for the introduction of targeted mutations. The technique is undoubtedly a powerful way to obtain large contiguously linked segments of genes.
We developed a bit of a love-hate relationship with it, though. While we used it to successfully execute construction schemes on many occasions, there have been a few times when we simply could not amplify particular gene segments despite what seemed to be Herculean efforts to do so, including the use of a multiplicity of enzyme systems, reaction adjuncts, cycling conditions, etc. In one case we narrowed it down to a several-hundred base-pair segment that, if included in the target sequence, could not be “crossed” by PCR, presumably due to a structural constraint. So while it is great when it works, there is no way to know a priori that a target sequence will be amplifiable.
In general, our preferred strategy to study deletions, duplications, and copy-number variations is through digital PCR. We use the OpenArray with 3,042 PCR chambers from Life Technologies. The accuracy in copy-number estimates with digital PCR far exceeds those with conventional qRT-PCR.
My laboratory has invented LATE-PCR and a set of complementary chemistries for sample preparation (Quantilyse and PurAmp), suppression of mispriming PrimeSafe™, and product detection and analysis over sequences up to several hundred nucleotides long (Lights-On/Lights-Off Probes). LATE-PCR and RT-LATE-PCR efficiently generate double-stranded DNA products and then automatically switch to efficient generation of single-stranded DNA products.
The lengths of these products can be from ~70 to ~600 nucleotides depending on their sequence composition and the specific needs of the assay. Lights-On/Lights-Off probes are paired probes in which a fluorescent signal is generated by the binding of the Lights-On probe and then is extinguished by binding of the Lights-Off probe to the adjacent sequence. Using this strategy, we are able to coat the target using just a single fluorescent color and the resulting signals from all of the probes generate a fluorescent signature (as a first derivative), which is characteristic of the specific target sequence.
Deletions, insertions, and duplications dramatically alter the fluorescent signature and are thus easily detected. Indeed, a single pair of Lights-On/Lights-Off probes can be used to scan for duplications or deletions of a particular sequence while simultaneously revealing the presence of sequence variants. Such an approach is well suited for closed-tube rapid diagnostics and gets around the need for sequencing the product. We call this approach Virtual Sequencing. Because reactions of this design are sensitive down to the single-molecule level, we envision their use for scanning for sequence changes in rapidly evolving RNA viruses such as HIV.
Micro-electro-mechanical systems (MEMS) technology has been used to adapt real-time PCR machines to lab-on-a-chip. What are the benefits and shortcomings of real time PCR chips?
Drs. Chittur and Tine
The primary advantages of real-time PCR chips are the speed of the systems, their high sensitivity, and the portability that is allowed by their size. The small scale also provides obvious economies by reducing the amount of reagents required, and hence the costs of running the assays. It is not hard to imagine that these characteristics will facilitate the usage of such systems to provide widespread access to qPCR-based assays in point-of-care healthcare settings such as physicians’ offices, hospitals, and mobile clinics.
Their portability will be particularly suited to applications in forensic science and bioweapon threat detection. Many of these applications would be most useful with crude samples, but the sensitivity of PCR to the presence of inhibitors will require a level of template purification that may prove to be restrictive.
While the small reaction volumes provide advantages for size, speed, and portability, they also limit the template concentrations that can be directly analyzed without the need for further manipulation. Another factor that has plagued microfluidics-based systems is the efficient mixing of the reaction mixture without introducing air bubbles.
The main advantages are lower cost per reaction and usually the much higher throughput offered by chip-based platforms. The primary disadvantages are the small sample volumes, which require pre-amplification that may introduce bias and are not applicable for all qPCR applications. In addition, specialized reagents and auxiliary equipment for loading, etc., are needed.
I refer to these systems as ultrahigh-throughput qPCR platforms. These technologies are capable of performing tens of thousands of qPCRs in the standard two-hour run time. qPCR lab-on-a-chip systems may ultimately replace conventional oligonucleotide microarrays due to the increased sensitivity of PCR.
Using this technology, it may be possible to profile the entire human transcriptome in a two-hour period. In addition to throughput, another advantage of these systems is the reduced reagent costs due to the small reaction volume (nanoliter range).
A shortcoming of these systems is that traditional pipetting will be replaced with more robotic types of loading. This may limit the use of ultrahigh throughput qPCR profiling to those comfortable with robotics or more dedicated users such as core facilities.
Also, these technologies may be more applicable to core facilities rather than individual users due to the sophistication of the technology.
An important application of PCR is molecular staging of cancers. The RT-PCR assay has the potential for generating data that is more informative than conventional histopathology. How is this approach superior to classical staging techniques, and what caveats must researchers keep in mind before using the RT-PCR assay for this application?
Drs. Chittur and Tine
Histopathology-based staging of cancers is very well developed and widely applied in the field. However, minimal residual disease in cancer patients may be undetectable using conventional methods due to the low abundance of disseminated tumor cells. This is an area where RT-PCR-based assays may be more acceptable. The advantage of this approach is that it provides a more precise, molecular basis for staging a tumor. Ideally, this will improve the accuracy of staging, which in turn will lead to the use of the most appropriate therapies and increase the chances for a positive outcome.
The sensitivity of PCR should allow the detection of aggressive variants within a tumor population that would be undetectable by more conventional approaches and would inform the choice of therapy. The main caveat is the need to reliably establish the correlation between the expression of a particular gene or variant with the malignant potential of the tumor. Issues related to sample collection, technical variations in the protocol, and primer design can affect the overall results. Contamination of sample with aerosols containing amplified DNA from previous PCR assays may also be a source of false-positive results.
However, a well-designed assay performed with all the standard precautions has the advantage of exquisite sensitivity and thus prognosis/diagnosis of cancer can be superior as compared to conventional methods.
It is very important to validate markers. Typically, markers are selected by screening, which results in a large number of leads, but most of them will be false positives. To identify truly significant and informative markers, large validation studies are required. This is particularly challenging in multivariate studies, where the correlated response of multiple markers is used to reflect disease.
Another challenge is sample complexity. Cancer is a polyclonal disease and a few rare cells with stem cell-like properties may determine the response to treatment and chances for survival. Furthermore, disseminated tumor cells in bone marrow and circulating tumor cells in peripheral blood that give rise to metastasis are more relevant for treatment decisions and prognosis than cells from the primary tumor that is removed surgically. To approach this, our team developed single-cell qRT-PCR profiling.
While I have no formal training in pathology, I have worked with clinical tumor specimens for a number of years. I am not sure that any molecular assay will surpass conventional histopathology in terms of detecting the presence or absence of diseases such as cancer.
Molecular staging may be useful to diagnose diseases at their very earliest stage, prior to the development of significant pathological changes. Molecular staging may be useful in situations where it is difficult to obtain tissue specimens but easier to collect biofluids.
My lab is currently developing assays for the early detection of pancreatic cancer. We believe that measuring specific microRNAs in pancreatic specimens will be beneficial in detecting pancreatic cancer at its earliest stages. For example, molecular staging using microRNA expression may allow the clinician to determine which patients with pancreatic cysts are candidates for more aggressive treatment.
Using qPCR to assay microRNA levels in pancreatic cystic fluid obtained by endoscopic ultrasound may provide useful molecular staging information in addition to or in place of conventional histopathology. From a clinical perspective, consistent collection of the specimens must be adhered to so that the RNA is not degraded prior to analysis.
Sample handling and RNA preparation are critical to the success of RNA analysis and have to be standardized.
Real-time PCR assays are widely employed in evaluating Helicobacter pylori culture and antibiotic susceptibility, particularly SNPs associated with clarithromycin and tetracycline resistance. Has PCR detection for diagnosis and drug resistance essentially replaced other methods, or are these alternatives able to offer significant advantages to the PCR approach?
Drs. Chittur and Tine
qPCR detection of H. pylori and identification of SNPs associated with drug resistance complement the classical approaches to diagnosis such as histological detection and culture-based agar diffusion assays to determine drug sensitivity. They are significantly more rapid than classical tests and allow detection from clinical specimens obtained by noninvasive means. There are some disadvantages, however.
If many mutations are found to confer resistance to a particular drug, then assay design can become cumbersome. In addition, the presence of a novel mutation would not be detected by these qPCR-based approaches. Most bacterial infections will consist of a mixture of genotypes. The ability of qPCR-based detection to identify the presence of resistant bacteria as a very minor component within a larger population of sensitive organisms (prior to treatment with an antibiotic that ends up expanding the population of resistant organisms, eventually resulting in treatment failure) would provide a significant advantage to this type of testing and increase its usage.
There are companies that specialize in multiplex qPCR for pathogen detection. Alternative platforms, however, often are as good as qPCR for this purpose. It is challenging to develop multiplex qPCR tests with high sensitivity and specificity for clinical samples.
It is my understanding that the traditional approach to detecting SNPs using techniques such as pulse field electrophoresis has been replaced by qPCR using fluorescence-based hydrolysis probes.
The qPCR technology would have several advantages over this traditional approach. Hydrolysis probes are designed to emit fluorescence following 100% hybridization to the target sequence. This makes them ideal for SNP detection. Also, the increased throughput and closed system of the qPCR are advantageous.
Lights-On/Lights-Off probes are ideally suited for distinguishing wild-type and drug-resistant strains. Indeed, we are currently finishing construction of an assay for M/XDR- TB, which can distinguish all of the significant mutations responsible for rifampicin resistance using a single color and in other colors can simultaneously detect and distinguish the major mutations for resistances to isoniazid, ethambutol, fluoroquinolones, and aminoglycosides.
This multiplex assay also distinguishes M. tuberculosis, M. bovis, and M. aftricanum (all MTBCs) from approximately 20 different species of NTM (nontuberculosis Mycobacteria).