PCR enables the amplification of a specific genetic sequence. In conventional PCR, a sample containing a dilute concentration of template DNA is mixed with a heat-stable polymerase enzyme, primers and deoxynucleoside triphosphates (dNTPs) in a suitable buffer.1 Through iterative cycles of heating and cooling, the DNA template is repeatedly denatured, followed by annealing of primer oligonucleotides and extension/elongation of complementary strands, leading to exponential amplification of the template DNA. The amplified product is typically detected using agarose gel electrophoresis, the end-point analysis to determine the presence of the target DNA in conventional PCR.
In real-time/quantitative PCR (qPCR), the amplified product is detected using fluorescence methods in real time as the reaction proceeds. For qPCR, the target sequence undergoes rounds of PCR cycles, exponentially increasing the quantity of DNA copies of the target sequence (amplicon). Fluorescently tagged DNA probes are designed to bind to the specific amplicon for a given PCR reaction. During amplification, a proportional amount of fluorescent reporter dye is released relative to the amount of the target amplicon. The quantity of target DNA present in the sample is extrapolated from the fluorescence signal emitted from a plasmid DNA reference or standard curve in this case, allowing for only relative quantification.
Digital PCR relies on statistical analysis2,3 to determine the absolute concentration of the target DNA independent of a plasmid DNA standard curve. The Droplet Digital™ PCR System partitions an appropriately diluted DNA sample into approximately 20,000 droplets using oil as an emulsifying agent. Ideally, each oil droplet contains zero or one target molecule — thus providing a binary, digital readout. During PCR amplification, the fluorescent reporter dye is released, indicating the presence of the target amplicon in the droplet. The fluorescence of the droplet is measured using flow cytometry, and the individual droplets are scored positive or negative based on the fluorescence signal. The distribution of positives to negatives in each sample (Poisson statistical analysis) yields absolute quantitation of the target sequence. In addition to ddPCR, other methods of PCR include digital PCR performed on a microfluidic chip or following separation onto microarrays.
The notable difference between qPCR and ddPCR is that in the latter there is no need to rely on numerous amplification cycles to determine the amount of template nucleic acid in each sample relative to a standard curve. For qPCR, delays in amplification (for example, due to secondary structure in DNA) can yield a false increase or decrease in target DNA quantification. Conversely, for ddPCR, the distribution of positive and negative droplets used for the statistical analysis provides an absolute quantification of the template. ddPCR also exhibits better accuracy, sensitivity and precision than qPCR. It is thus ideal for the detection of small quantities of target DNA or analyses that require fine resolution, such as rare sequence detection, copy number variation (CNV) analysis and gene expression analysis of rare targets.
The value of absolute quantification afforded by ddPCR cannot be overemphasized. The accuracy and precision of the results obtained via relative quantification using qPCR are directly correlated with the standard/reference curves used for these experiments. Because there is often variability in the design and quality of standard curves, the accuracy of the titer results will depend on how effectively the standard curve has been optimized.
This variability in the standard curves used for qPCR can in part be attributed to the different architectures of the relevant DNA species — circular, supercoiled versus linear DNA. Each form of DNA used for the standard has different PCR efficiencies itself, and may have efficiencies different than the target amplicon if it has a different structure. Therefore, the use of unoptimized standard curves will result in inaccurate quantification, which can serve as a source of significant lab-to-lab variation in qPCR results.4
The coefficient of variation (<10%) associated with ddPCR is lower than those achieved using qPCR. This provides more reliable titers of the viral vector for determining dosing, which is critical when performing dose-ranging studies for toxicology and in the clinic.
Sample partitioning in the ddPCR system allows the sensitive, specific detection of single or multiple template molecules. These combined qualities make ddPCR suitable for a wider range of applications than conventional PCR and qPCR, including liquid biopsy, copy number variation (CNV), rare sequence detection, gene expression and miRNA analysis, single-cell analysis, pathogen detection and next-generation sequencing library analysis.
Gene and cell therapies utilize a variety of viral vectors for gene transfer, including adeno-associated viruses (AAV), lentiviruses, retroviruses, herpes viruses, adenoviruses and others. AAV, adenovirus and HSV vectors contain DNA, while lenti- and retroviral vectors contain RNA. PCR plays a critical role in many analytical methods used to characterize and quantify these viral vectors during production, product release and stability testing.
Vector genome titering involves quantification of the viral vector genome and has largely been used for DNA-containing AAV vectors, but also increasingly for other vector types. Infectious titering is used to determine the infectivity of viral vectors. To perform infectious titer assays, dilutions of the viral vector are “sprinkled” on cells in culture, and the infectious virus can be detected using qPCR as an endpoint readout. qPCR serves as a surrogate for manual observation of cytotoxic effects, such as counting the number of dead cells/plaques.
Residual DNA impurity testing is performed to quantify residual plasmid, host-cell and unpackaged viral DNA potentially present in the viral vector product. The specific test(s) will be determined by the type of viral vector and the process used to manufacture it, as well as the specific amplicons. Residual unpackaged viral DNA can be quantified by ddPCR, and helper viruses harboring the therapeutic vector genome can be detected if a helper virus is used in the production process.
qPCR is also applied in assays designed to detect replication-competent viruses that have the potential to be generated during the production of viral vector products. To ensure patient safety, most viral vectors are engineered to eliminate their ability to replicate once administered in the human body. In the design of the therapeutic vector, the viral genes required for replication are replaced with therapeutic genes in the therapeutic viral vectors and supplied in trans on a separate molecule. The potential for recombination is monitored using assays to detect replication-competent virus. A qPCR assay is performed as an end-point readout following an infectious amplification to confirm the absence of replication-competent virus in the product.
The preparation of the vector sample includes steps to reduce/eliminate external, unpackaged DNA and the liberation of the viral vector genome from the viral particle. If the viral vector genome is RNA, a reverse transcriptase step may be needed. Proper sample preparation is crucial to the success of ddPCR methods. Since ddPCR is an assay based on statistical analysis of target molecules partitioned into droplets, production of a homogenous solution before droplet formation is essential to ensure even distribution of the target molecules in all droplets. This is important, as the degree of homogeneity impacts the ratio of positive and negative droplets, which, in turn, affects the quantification of the target molecule.
Overall, the higher precision and accuracy afforded by PCR, including ddPCR, is largely driven by the amplicon design and the efficiency of the amplification. The amplicon is customized to meet the specific requirements of a given assay and the viral vector involved. The design of a ddPCR assay is similar to a qPCR assay, so development of the amplicon when converting from qPCR to a ddPCR method typically involves slight modifications to achieve enhanced performance, but the primers and probes are generally very similar.
The keys to good amplicon design include avoiding secondary structures, selecting the right amplicon size and identifying the proper sequence and placement for the primer and probe. Shorter amplicons are preferred to minimize the formation of secondary structures and to maximize efficiency. For end-point PCR analyses, the amplicon must complete the reaction and achieve elongation. ddPCR is less affected by secondary structures than qPCR. In qPCR, secondary structure of the DNA target delays amplification and impacts quantification; in ddPCR — although a potential decrease in the resolution of positive and negative droplets may be observed — it remains possible to differentiate between negative droplets and droplets with intermediate fluorescence (often referred to as rain), allowing accurate quantification.
It is critical to verify that primer–probe sets perform when used under different conditions and are appropriate for the intended purpose. Primers must distinguish the genetic payload (target DNA or cDNA) from any endogenous genes that are present in the production cell line. Vectors typically express only the exonic regions of their therapeutic genes, while the endogenous cellular genes contain both exons and introns. Thus, PCR amplicons in vectors are typically designed to traverse exon–exon boundaries, which prevents detection of endogenous sequences containing introns. Primers used in PCR analyses can also be designed to target promoters present in the vector genome, which typically are heterologous sequences not naturally present in production cell lines.
Brammer Bio uses qPCR extensively, though it requires standard/reference curves and the ability to control the associated variability. We have begun converting our qPCR methods to ddPCR analyses, starting with vector genome titer assays for AAV vectors. We are currently developing additional methods, including vector genome titer analyses for RNA-containing vectors (retroviral, lentiviral) and other vector genome configurations, residual DNA impurity analyses and methods for the detection of replication-competent vectors.
We have adopted the Bio-Rad platform — the QX200 Droplet Digital PCR — which includes the automated droplet generator, thermocycler and droplet reader. Our use of the automated droplet generator has proven to ensure better repeatability for droplet generation compared with manual systems.
The specific steps involved in preparing a sample for ddPCR analysis depend on the type of viral vector and the purpose of the assay (e.g., impurity analysis, vector genome titer). For encapsidated viruses, it is important to ensure that nothing outside of the virion can interfere with or skew the results before release of the viral DNA/RNA from the virion. We have developed effective procedures for the removal of exogenous DNA and RNA before liberating the viral vector genome from the virus. This step is followed by reverse RNA transcription for lenti- and retroviral vectors or restriction digestion for special configurations of AAV genomes. These methods are routine for Brammer and are designed to ensure enhanced accuracy of our ddPCR methods.
We have established the appropriate parameters on the basis of the viral vector characteristics relevant to ddPCR. At Brammer Bio, we have focused on optimizing these steps, as well as the PCR conditions for each assay (e.g., primer/probe concentration, annealing temperature, limits of detection and quantification).
Other optimized parameters include the resolution of the digital assay (separation between positive and negative droplets), the threshold determination, and the decrease of rain in the analysis. We have also built our capability to recognize the optimum window of quantification. Samples for ddPCR must be diluted to an appropriate range for distribution of positive and negative droplets. If a sample is too concentrated or too dilute, the analysis will not provide accurate results. Although ddPCR has a slightly smaller dynamic range than qPCR, Brammer has developed dilution strategies that enable us to use ddPCR to monitor the amount of viral vector during the production and purification process, as well as to characterize the final product. We have also focused on simplifying assays where possible. Most residual host-cell DNA analyses require purification of DNA using purification columns before qPCR. The DNA purification step often results in variable recovery of the host cell DNA, which prevents accurate determination of the residual host-cell DNA concentration. We have optimized sample preparation before ddPCR to more accurately determine the amount of host-cell DNA.
Dr. Snyder was the founder of Florida Biologix, which was spun out of the University of Florida in 2015 and merged to create Brammer Bio in 2016. Dr. Snyder has been investigating virus biology, vector development, cGMP manufacturing and analytical technologies, and viral vector–mediated gene transfer for over 32 years. Dr. Snyder received his doctoral degree in microbiology from the State University of New York at Stony Brook and obtained his BA in biology from Washington University in St. Louis.