|Home | About | Journals | Submit | Contact Us | Français|
Human rhinoviruses (HRV) comprise 3 species representing more than 150 genotypes. As an important human respiratory pathogen, molecular detection is an indispensable tool for diagnosis and surveillance. However, the sequence diversity of HRV genotypes poses challenges for developing robust molecular methods that detect all genotypes with equal efficiencies. This study compares the accuracies of reverse transcription-quantitative PCR (RT-qPCR) and reverse transcription-digital PCR (RT-dPCR) for quantifying HRV RNA using genotype-specific primers and probes and a consensus primer/probe set targeting the 5′ noncoding region of HRV. When using consensus primers and probes for the quantification of HRV, RT-dPCR outperformed RT-qPCR by consistently and accurately quantifying HRV RNAs across more genotype groups, despite the presence of up to 2 target-sequence mismatches within the primer or probe binding region. Because it does not rely on amplification efficiency, which can be affected by sequence mismatches in primer/probe binding regions, RT-dPCR may be the optimal molecular method for future HRV quantification studies and for quantitating other viruses with high sequence diversity.
Human rhinoviruses (HRV) are important human respiratory pathogens that are small positive-sense RNA viruses within the family Picornaviridae. There are more than 150 genotypes of HRV that have been recognized within species A, B, and C of the genus Enterovirus (1). Molecular assays, such as reverse-transcription PCR (RT-PCR), are the most useful methods for detecting HRV in clinical samples (2,–4). Most HRV RT-PCR assays target the conserved 5′ noncoding region (NCR), which exhibits the greatest sequence homology among the HRV genotypes. However, even in the 5′ NCR, consensus primer and probe sets must be designed with degenerate and modified bases or multiple oligonucleotides to amplify all HRV genotypes (5,–7).
Although qualitative detection of HRV by RT-PCR is currently sufficient for determining HRV infection, accurate quantification of HRV RNA in clinical samples is needed for studies associating HRV viral load with viral transmission and with patient symptoms and outcomes. Viral load studies of other respiratory viruses have shown that a correlation exists between viral load and disease severity (8,–10). Accurate quantification of HRV will also be required for evaluating the performances of future antiviral drugs. Real-time RT-PCR assays, when accompanied by the amplification of standard curves (RT-qPCR), can be used to quantify the number of viral copies in a sample. However, RT-qPCR assays using quantification with a consensus HRV primer and probe set may not give accurate results for all genotypes due to amplification inefficiencies caused by base mismatches between the primer and probe sequences and the specific viral sequences (11). Digital PCR (dPCR), which quantifies nucleic acids independently of a standard curve and is less affected by PCR efficiency (12, 13), may perform better for quantifying HRV RNA by RT-PCR.
The aim of this study was to investigate the best method for quantifying HRV RNA in clinical specimens. The efficiency and sensitivity of a consensus RT-qPCR assay were compared with those of genotype-specific RT-qPCR assays to determine if consensus RT-qPCR accurately quantifies all genotypes of HRV, including those with sequence differences in the primer and probe binding regions of the consensus assay. Also, to determine if consensus RT-dPCR is more accurate for quantifying all HRV genotypes than is consensus RT-qPCR, we compared quantification results from RT-dPCR with those from RT-qPCR for known amounts of RNA from different HRV genotypes using a consensus set of HRV primers and probe targeting the 5′ NCR.
HRV RNA transcripts from each of the 16 genotype sequence groups were tested with the genotype-specific primer/probe sets and the consensus primer/probe set by RT-qPCR and RT-dPCR. When the results from type-specific and consensus primer/probe sets were compared in RT-qPCR assays of the transcripts, the genotype-specific assays were more sensitive for HRV detection than were the consensus assays for 5 of the 16 HRV genotype groups (F, I, N, R, and U) using a 10-fold or greater difference in sensitivity as the threshold (Table 1 and Fig. 1). The consensus and genotype-specific RT-qPCR assays showed equivalent sensitivities for ten genotype groups, while the consensus assay was more sensitive than the genotype-specific assay for 1 genotype group (J) (Table 1). However, the RT-dPCR genotype-specific and consensus primer/probe sets performed similarly on all but one of the 16 HRV genotype groups (group N) (Fig. 2). The number of sequence mismatches between the consensus primers and probe and the 16 HRV genotypes ranged from 0 to 2 (Fig. 3 and Table 1). Single mismatches in the forward primer binding region resulted in losses of sensitivity up to 2 logs, depending on the location of the mismatch.
We determined the HRV viral loads in RNA extracted from clinical samples positive for 55 different HRV genotypes by RT-qPCR using genotype-specific primers and probe and compared the results with the viral loads determined when amplified by consensus RT-qPCR and RT-dPCR. For clinical samples containing HRV genotypes with sequences that closely matched the consensus primer and probe sequences (e.g., genotype groups A, B, C, and D), quantifications of HRV RNA by RT-qPCR and RT-dPCR performed with the consensus primers and probe showed good agreement with the viral load determined by genotype-specific RT-qPCR (Fig. 4). In contrast, for some HRV genotypes with 1 to 2 mismatches within the consensus primer and probe sequences, RT-qPCR performed using these primers and probe underestimated the viral load by up to 2 logs (Fig. 4). Strikingly, RT-dPCR with the consensus primer/probe set provided more accurate quantifications than did consensus RT-qPCR on multiple samples, particularly within genotype groups I, R, and U (Fig. 4). Four samples within genotype groups I, R, and U showed a ≥1 log improvement in quantitation by RT-dPCR compared with that by RT-qPCR. The viral loads of clinical samples ranged from 0.85 log10 to 7.6 log10 copies with a median of 4.47 log10 copies, and the improvement in quantitation was observed on samples that had viral loads ranging from 0.85 log10 to 5.77 log10 copies.
Although infections with HRV can be determined using qualitative RT-PCR, an accurate quantification method is needed to determine if HRV viral load is associated with viral transmission and pathogenicity and with patient symptoms and outcomes, as has been shown for other respiratory viruses. HRV viral load determinations may also be important for patient management, especially in asymptomatic patients who test positive for HRV at low levels. More importantly, accurate HRV viral load assessments will be necessary for evaluating the performances of potential HRV antiviral drugs. The quantification of HRV by RT-qPCR assays using type-specific primers and probes is accurate. However, these assays must be used on samples for which the HRV genotype is known so that the appropriate primers and probe and standard curve can be included. In this study, we investigated the accuracy of quantifying HRV by RT-qPCR when consensus primers and probes were used and compared the results to those from RT-dPCR.
We showed that the consensus RT-qPCR accurately quantified many HRV genotypes. However, these data also demonstrated that RT-qPCR using consensus primers and probe did not accurately quantify all genotypes of HRV, especially HRV species C, due to the suboptimal amplification of genotypes with sequences that did not exactly match those of the primers and probe. When using consensus primers and probe for quantifying HRV, RT-dPCR outperformed RT-qPCR by consistently and accurately quantifying HRV RNAs across more genotypes. Although we tested a large number of HRV genotypes representing most of the types with sequence mismatches in the consensus primer and probe binding regions, not all HRV genotypes were tested, especially the HRV species C genotypes. Some did not have 5′ NCR sequences available, and for some genotypes, positive specimens were not available for testing. However, the data from 16 genotype sequence groups that represent sequences of 128 genotypes indicated that RT-dPCR quantification of HRV RNA will be useful in future HRV viral quantification studies. The typical workflow for RT-dPCR requires more hands-on technician time than do current RT-qPCR clinical workflows. However, the tradeoff for greater accuracy with only a single consensus primer/probe set offsets the extra few hours in run-time. A final limitation of RT-dPCR in comparison with RT-qPCR is the more limited dynamic range (104 for RT-qPCR compared with 108 for RT-qPCR). However, we did not need to dilute any of the 55 clinical samples tested here to obtain the concentration ranges required for the RT-dPCR assay, so this is not expected to be a major issue.
In this study, we use type-specific and consensus primers and probes to test known copy numbers of a large number of HRV genotypes that represent most of the sequence variation found in the consensus primer and probe binding region. In comparison, other reports on the quantification of HRV used only consensus RT-qPCR assays, quantified from standard curves generated from a single HRV genotype, and tested a limited number of HRV genotypes (14,–16). Nevertheless, our findings confirm the results from some of these studies and show that assay quantification ability is linked to target variability and that accurate HRV quantification by a single RT-qPCR assay is not feasible for all HRV genotypes (17, 18).
The sequence diversity of HRV and the availability of well-characterized 5′ NCR target sequences for over 50 genotypes provide a good model for testing the hypothesis that dPCR is more recalcitrant to sequence mismatches within primers and probe than is qPCR. It has been theorized that dPCR is less susceptible to amplification inefficiency caused by primer/probe sequence mismatches because the quantification is derived from a PCR that cycles to endpoint rather than from an amplification curve as in qPCR (13, 19). Until now, there have been limited empirical data to support this hypothesis. However, these HRV data clearly support the aforementioned hypothesis and demonstrate dPCR's advantage in cases of primer/probe sequence mismatch.
However, it should be noted that dPCR did not overcome all sequence mismatch-induced amplification inefficiencies, as evidenced by genotype sequence group N (Fig. 2), which has a single mismatch near the middle of the forward primer. Clearly, the number of mismatches is not a perfect predictor of the amplification inhibition. Other genotype sequence groups with more mismatches than genotype group N did not show amplification inhibition by RT-dPCR. It is likely that the effect of a sequence mismatch on amplification is highly context dependent, given that the 5′ NCR region of HRV harbors significant RNA secondary structure (20). Nonetheless, dPCR is a better alternative to qPCR on templates known to have significant sequence diversity that cannot be avoided during primer and probe design. dPCR should be considered the optimal molecular method for quantifying HRV in clinical specimens and may be applicable to other viruses with high sequence diversity, such as HIV and hepatitis B virus (HBV).
Type-specific and consensus primers and probes were used in RT-qPCR and RT-dPCR assays for quantifying known genotypes of HRV, including RNA transcripts of specific HRV amplicons and clinical specimens containing whole virus. For consensus RT-PCR, a previously designed HRV primer and probe set was used (5): forward primer, CPY+AGCC+TGCGTGGY; reverse primer, GAAACACGGACACCCAAAGTA; probe, 5′-FAM-TCCTCCGGCCCCTGAATGYGGC-BHQ. The 5′ NCR region sequences for all species A and B HRV and for 43 of 55 species C HRV were aligned and placed into 22 groups according to their sequence homologies to the consensus primers and probe (Table 2 and Fig. 3). For type-specific RT-PCR, 13 primer and probe sets were designed to match the specific sequences within the consensus primer and probe binding regions for 16 of the 22 HRV genotype groups (Table 2). Samples containing representative HRV genotypes from the other 6 genotype groups were not available for testing and therefore were not included in our analysis.
The 10 different type-specific forward primers were synthesized with T7 polymerase promoter sequences on the 5′ ends. RNA extracted from samples positive for a representative HRV genotype from each of the 16 genotype groups, including 14 serotyped HRV culture isolates and 2 sequenced clinical samples (groups R and U) (Table 2), was reverse transcribed into cDNA and amplified by conventional PCR using the type-specific T7 polymerase promoter-labeled forward primer and type-specific reverse primer. The purified type-specific amplicons were then transcribed into RNA (T7 Megashort script kit; Life Technologies). The type-specific RNA transcripts were purified and quantified as previously described (21) and used to generate standard curves for RT-qPCR. Dilutions of the 16 transcripts (10 to 1e7 copies/reaction for RT-qPCR and 1e2 to 1e6 copies/reaction for RT-dPCR) were amplified using the genotype-specific primers and probes and the consensus primers and probe in one-step RT-qPCR (AgPath one-step RT-PCR; Life Technologies) and one-step RT-dPCR (One-step RT-ddPCR advanced kit; Bio-Rad) assays. Different ranges of quantified viral transcript were used in qPCR versus dPCR to account for the different dynamic ranges of the two assay platforms. RT-qPCR was performed as previously described (5) except that the PCR annealing temperature was lowered to 50°C or 55°C when using the type-specific primers. RT-dPCR was performed as previously described on the QX100 droplet digital PCR instrument (Bio-Rad Laboratories, Hercules, CA), with One-step RT-ddPCR advanced mix replacing the ddPCR Supermix for probes (22, 23) and with the following reaction conditions: 50°C for 60 min, 95°C for 10 min, 40 cycles of 95°C for 30 s and 60°C for 1 min, followed by 98°C for 10 min. The forward and reverse primers were at final concentrations of 500 nM and the probe was at a final concentration of 100 nM. The lower limits of detection (LLOD) were determined for the RT-qPCR consensus and genotype-specific assays by running 8 to 16 PCR replicates of serial dilutions in the lower range of transcripts (10 to 10,000 copies/reaction) and determining a 95% cutoff by probit analysis.
For RT-qPCR and RT-dPCR, HRV RNA was extracted from 60 clinical samples (52 swabs and 8 culture isolates) positive for 55 known HRV genotypes representing the 16 genotype groups (identified by serotyping or sequencing and including 29 species A, 12 species B, and 14 species C) with the Total nucleic acid high performance kit on a MagnaPure LC 2.0 (Roche Diagnostics, Indianapolis, IN). HRV RNA was quantified by RT-qPCR using genotype-specific primers and probes and genotype-specific standard curves. The quantities of HRV RNA in each sample were then compared after quantification by RT-qPCR and RT-dPCR using the consensus primers and probe. The accuracy of RT-qPCR compared with that of RT-dPCR was assessed by determining which assay gave values within at least 1 log10 of the expected value determined by genotype-specific RT-qPCR.
R.H.S. has consulted for Bio-Rad Laboratories on projects unrelated to this article.