Even experienced scientists using qPCR for gene expression studies may make unwise assumptions regarding the quality and performance of the RNA template, and the amplification protocol that can lead to inaccurate or wrongly interpreted results.
Tania Nolan, Ph.D., business development manager at Sigma-Genosys (www.sigma.com), emphasizes the importance of performing quality control and optimization experiments, especially for real-time qPCR.
The company’s assay development work stemmed from the growing interest in profiling mRNA from archived biological samples, typically formalin-fixed, paraffin-embedded (FFPE) samples, and correlating the gene expression data with established clinical data sets.
“The RNA isolated from FFPE-archived material is highly degraded, and it was unclear how such degradation would affect data from such samples,” said Dr. Nolan. She set out to devise a high-throughput method for evaluating the integrity of the RNA templates. An additional aim was to produce a technique that could be used when only small amounts of sample are available, such as samples from laser-capture microdissected biopsies. A detailed analysis of these samples would ensure that RT-qPCR is performed on samples of equivalent integrity. Using the 3´:5´ assay, she demonstrated that the state of the RNA template can affect the outcome of RT-qPCR and the perceived results of gene expression studies.
The method begins with a target RNA transcript and reverse transcription using oligo-dT priming. The reverse transcription then starts from the polyA tail and moves from 3´ to 5´. Two sets of primers target the 3´ and 5´ areas of the specific cDNA template, which are amplified either in separate reactions or in a duplex assay using amplicon-specific probes.
Target quantification is measured using SYBR Green I dye. If the RNA is intact, then quantifying the 5´ and 3´ ends of the cDNA generated from the same target will yield the same quantity of target. In contrast, if the RNA is degraded somewhere between the 5´ and 3´ assay sites, then the reverse transcriptase will stall at the break point and the assay will yield more 3´ cDNA.
Dr. Nolan has observed that not all transcripts in a sample degrade at the same rate. In fact, transcripts appear to degrade at different rates in different tissues. Ongoing work at Queen Mary University of London, together with Sigma-Genosys, aims to determine which targets, and how many, need to be studied before it would be safe to assume whether or not a particular RNA sample is intact.
Another quality control issue relates to the risk of contaminants in the sample that can inhibit the enzymatic reactions required for RT-qPCR. Environmental samples or crude biological preparations, for example, may contain contaminants that inhibit PCR, resulting in false negative results. It is not safe to assume, though, that if a sample contains an inhibitor it will always inhibit all qPCR assays to the same degree. Therefore, noted Dr. Nolan, it is not sufficient to correct for inhibitors by quantifying a reference gene.
To identify whether a sample contains factors that may inhibit PCR, Dr. Nolan and colleagues developed the SPUD assay. This involves running a control PCR using an artificial amplicon (from the potato genome) in parallel with an identical PCR reaction that contains both the artificial amplicon and the sample. If the sample contains an inhibitor, it will take more cycles for the fluorescence signal to rise above the threshold background fluorescence (a parameter known as Ct, or threshold cycle). Therefore, in the presence of inhibiting factors, the Ct will shift higher.
Interestingly, an inhibitory factor may inhibit the 3´ PCR assay and the 5´ PCR assay to different degrees. While the SPUD assay can identify the presence of an inhibitor, the results of the 3´:5´ assay may also reflect the variable effects of the inhibitor, resulting in either more 5´ sequences or more 3´ sequences. Sigma-Genosys is designing a set of reference assays that will allow for a comparison of the amount of 5´ and 3´ sequences compared to a standard curve.
“You cannot use normalization to get around the problem of RNA degradation” or as an alternative to evaluating the quality and integrity of an RNA sample, concluded Dr. Nolan. She recommends running both the 3´:5´ and the SPUD assay on every RNA sample as a quality control measure.