This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cook, L.
Right arrow Articles by Jerome, K. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Cook, L.
Right arrow Articles by Jerome, K. R.

 Previous Article  |  Next Article 

Journal of Clinical Microbiology, November 2006, p. 4149-4156, Vol. 44, No. 11
0095-1137/06/$08.00+0     doi:10.1128/JCM.01230-06
Copyright © 2006, American Society for Microbiology. All Rights Reserved.

Multiplex Real-Time Reverse Transcription-PCR Assay for Determination of Hepatitis C Virus Genotypes{triangledown}

Linda Cook,1 KaWing Sullivan,1 Elizabeth M. Krantz,1 Arthur Bagabag,1 and Keith R. Jerome1,2*

Department of Laboratory Medicine, University of Washington Medical Center, Seattle, Washington 98195,1 Program in Infectious Diseases, Fred Hutchinson Cancer Research Center, Seattle, Washington 981092

Received 15 June 2006/ Returned for modification 29 August 2006/ Accepted 6 September 2006


arrow
ABSTRACT
 
A variety of methods have been used to determine hepatitis C virus (HCV) genotypes. Because therapeutic decisions for chronic HCV-related hepatitis are made on the basis of genotype, it is important that genotype be accurately determined by clinical laboratories. Existing methods are often subjective, inaccurate, manual, time-consuming, and contamination prone. We therefore evaluated real-time reverse transcription-PCR (RT-PCR) reagents that have recently become commercially available (Abbott HCV Genotype ASR). The assay developed by our laboratory starts with purified RNA and can be performed in 4 to 5 h. An initial evaluation of 479 samples was done with a restriction fragment length polymorphism (RFLP) method and the RT-PCR assay, and discrepant samples were sequenced. An additional 1,200 samples were then tested, and data from all assays were used to evaluate the efficiency and specificity of each genotype-specific reaction. Good correlation between results by the two methods was seen. Discrepant samples included those indeterminate by the RT-PCR assay (n = 110) and a subset that were incorrectly called 2a by the RFLP method (n = 75). The real-time RT-PCR assay performed well with genotype 1, 2, and 3 samples. Inadequate numbers of samples were available to evaluate fully genotypes 4, 5, and 6. Analysis of each primer-probe set demonstrated that weak cross-reactive amplifications were common but usually did not interfere with the genotype determination. However, in about 1% of samples, two or more genotypes amplified at roughly equivalent amounts. Further studies are necessary to determine whether these mixed-genotype samples are true mixtures or a reflection of occasional cross-reactive amplifications.


arrow
INTRODUCTION
 
More than 60 genotypes and subtypes of hepatitis C virus (HCV) are distributed worldwide. Because therapeutic response rates differ by genotype, accurate and efficient genotype determination is critical to ensure correct treatment decisions.

Initially, genotypes 1 and 2 were identified by the sequencing of multiple samples in the HCV core region (6, 16, 23). Analysis from samples worldwide and sequencing of additional genome areas including the 5' untranslated region (UTR), envelope, and NS5 regions identified genotypes 3, 4, 5, and 6. Sequence variation between genotypes, subtypes, and individual strains is greatest in NS5, less in the envelope and core, and least in the 5' UTR.

Direct sequencing is the most accurate method for HCV genotyping. However, many other methods have been used because of the expense and technical difficulties of direct sequencing. The most frequently used methods in clinical laboratories are LIPA (line probe assay) and sequencing of the 5' UTR (both from Bayer Diagnostics). Both assays require two steps: generation of the PCR amplicon and then evaluation of the amplicon by either hybridization or sequencing. An additional disadvantage is the use of the 5' UTR, which is less informative for genotyping than are other, more variable regions of HCV.

In contrast, type-specific PCR assays require only a single amplification step and thus should be technically simpler. The first type-specific genotyping assay was described by Okamoto et al. (24). The assay utilized a first-round amplification of a large section of the core region, followed by a set of second nested amplification reactions each using an identical 5' primer but one of four subtype-specific primers (for genotypes 1a, 1b, 2a, and 2b). The method of Okamoto et al. detected the four different-sized amplicons by gel electrophoresis. A similar method was described by Chayama et al. (4) in the NS5 region. Larger-scale studies by Okamoto and others with their method demonstrated a significant lack of specificity between the type I (1a) and II (1b) reactions. Improved versions of the assays were described by Okamoto et al. (25), Holland et al. (15), Forns et al. (11), and Ohno et al. (22), which improved specificity and expanded the primer-probe sets to include the newly identified genotypes 4, 5, and 6. Several studies compared these type-specific assays to a variety of other genotyping methods (11, 13, 17, 19). In each of these studies, results from the type-specific assays had good agreement with other methods but had somewhat high rates of mixed samples and samples that did not amplify. Recently, several methods have been described that utilize single-step real-time reverse transcription-PCR (RT-PCR) with target-specific TaqMan (20, 28) or LightCycler (3, 31) probes or SYBR green detection (12), further streamlining the method. Some of these real-time methods were designed only to distinguish between genotype 1 and non-1 types, and none have been evaluated with large numbers of samples.

Reagents for a multicolor three-tube real-time RT-PCR assay for HCV genotyping have recently become commercially available (Abbott HCV Genotype ASR). We describe here the results from a large study designed to evaluate the performance of the real-time RT-PCR HCV genotype assay compared to our existing restriction fragment length polymorphism (RFLP) and core sequencing methods.


arrow
MATERIALS AND METHODS
 
Sample selection. All samples were submitted to the University of Washington molecular virology laboratory for clinical HCV genotype testing. Three groups of samples were used. Group 1 samples (n = 276) consisted of two subgroups that were specifically selected based on their previous RFLP genotype result and had been stored at –70°C for up to 3 years. The first 174 specimens of group 1 had approximately equal numbers of each of the common genotypes (1a, 2a, 2b, and 3a) and included all available genotype 4, 5, and 6 samples. In addition, some samples with unusual RFLP patterns of various types were included. Based on initial comparison data, an additional 102 samples, representing all available RFLP genotype 2a samples tested during the previous year, were added to produce the 276 total samples in group 1. All samples in this group were tested with the real-time assay. Group 2 samples (n = 203) were submitted for routine clinical typing and tested in parallel using the same RNA on the same day using both the RFLP and real-time methods. Group 3 samples (n = 1,200) were clinical specimens submitted to the laboratory over 9 months and tested only by the real-time assay.

Sample extraction and RNA purification. Serum or EDTA plasma samples were extracted using the Roche MagNA Pure LC instrument with the MagNA Pure LC Total Nucleic Acid Isolation kit (large volume), according to the manufacturer's instructions. The initial sample volume was 1.0 ml, and the elution volume was 80 µl (12.5x concentrated). For the real-time assay sensitivity experiment, serial 1:4 dilutions of serum in diethyl pyrocarbonate-treated water were made and then each dilution was extracted individually.

RFLP and Abbott real-time assay analysis. The RFLP assay was performed on amplicon product from the 5' UTR (8). The real-time assay consisted of three reactions, each of which contained three probes. The first reaction contained internal control primers-probes designed to detect all HCV genotypes (total HCV quantitation) and primers-probes for genotypes 1a and 1b. The second reaction contained primers-probes for genotypes 2a, 2b, and 3. The third reaction contained primers-probes for genotypes 4, 5, and 6. The genotype 1a and 1b reactions amplified sequences within the HCV NS5 region, while the other reactions amplified sequences within the 5' UTR. RT-PCR amplification was done for 50 cycles on an ABI 7000 instrument, and the raw data file was analyzed using Sequence Genotyping Software, v2.0 (Celera Diagnostics, Alameda, CA). Initially, the presence of two or more genotype reactions having cycle thresholds (CTs) within three cycles of each other was taken as evidence of mixed infection. Samples positive with the total HCV reaction but negative for all genotype-specific reactions were considered indeterminate. Raw amplification data were exported into Excel spreadsheets and analyzed using SAS statistical software (SAS Institute Inc.) and GraphPad Prism statistical/graphical software, v3.03 (GraphPad Software, Inc.). Genotype-specific amplification efficiencies (E) were calculated from the CT values of serial dilutions (E = 10–slope – 1) (29). Theoretically perfect replication efficiency for a PCR using this calculation would be 100%. For some highly efficient reactions the calculated amplification efficiency may slightly exceed 100%, presumably due to minor variations in serial dilutions.

HCV sequencing. Sequencing was performed on most samples by amplifying the core region by nested RT-PCR using primer sets described by Bukh et al. (2). The resulting 350-bp amplicon was sequenced using the ABI Big Dye terminator kit and an ABI 3730 sequencer. Sequence analysis was done with Seqscape software, and the genotype was assigned by matching with a library of 186 HCV core region sequences. The genotype assignment for each sequence in the library was confirmed by comparison with the Los Alamos database and by the generation of a phylogenetic tree using PHYLIP. A small subset of the samples were sent to Abbott Molecular's Research and Development group and sequenced in the NS5b region.

Statistical methods. Generalized estimating equations were used to test whether mean differences between CTs for the total HCV quantitation reaction (quantitative CT) and the individual genotype-specific reactions (genotype CT) were significantly different from zero. These models account for any correlation among samples belonging to the same laboratory run. This method was also used to compare the mean differences for the genotype 1b test to the mean differences for the other genotype-specific tests. We adjusted for multiple comparisons using the Bonferroni correction (10). The relative efficiencies of each genotype were calculated by subtracting the genotype-specific CT from the total HCV CT for samples with a single genotype amplification only. In an effort to compensate for differing relative efficiencies when comparing genotype-specific CT values, adjustments to the genotype CT based on these relative efficiencies were considered. Depending on the genotype, either a constant adjustment, no adjustment, or an adjustment based on the relationship between the relative efficiency and the genotype CT (using generalized estimating equations) was explored. Only samples with a single genotype amplified were used in developing these modifications. Sufficient numbers of samples were present for the calculations to be valid for the genotype 1a, 1b, and 3 groups across the entire range of CT values seen. However, an insufficient number of samples of low quantity of genotypes 2a, 2b, and 4 were available, so adjustments could not be accurately calculated in this range and thus were restricted to samples with greater quantities. Adjustments were then applied to samples with multiple genotype amplifications but only for CT data within the range of the data used to generate the adjustments mentioned above. This last restriction impacted genotype 2a and 4 samples most heavily, where more than half of the signals were out of range and thus could not be adjusted.


arrow
RESULTS
 
Performance characteristics of the real-time RT-PCR HCV genotyping assay. (i) Sensitivity and reproducibility. In order to measure the analytical sensitivity of the real-time assay, one sample for each genotype (1a, 1b, 2a, 2b, 3, 4, and 6) was serially diluted and analyzed by the total HCV and genotype-specific reactions. The lower limits of detection for genotypes 2a, 2b, 3, 4, and 6 in the genotype-specific reactions ranged from 92 to 317 IU/ml. In contrast, the genotype 1a and 1b samples were amplified in the genotype-specific reactions only when the quantity of HCV was greater than 1,500 IU/ml, indicating a substantially poorer sensitivity. The assay efficiency was highest for genotypes 2a and 6 (104.6% and 104.4%, respectively); slightly lower for genotypes 2b, 3, and 4 (85.6%, 89.2%, and 87.3%, respectively); and much worse for genotypes 1a and 1b (74.5% and 45.5%, respectively). The total HCV reaction showed similar sensitivities for all genotypes. The efficiency of the total HCV reaction for the seven genotypes averaged 90.9% (range, 81.3 to 107.8%).

To evaluate the reproducibility of the assay, a pooled positive-control serum, containing a mixture of genotypes 1a, 1b, 2a, 2b, and 3, was tested in 64 sequential runs. All 64 runs had detectable amplification signals for the total HCV, 1a, 2b, and 1b reactions. The mean CTs for the total HCV, 1a, 2b, and 1b reactions were 27.2, 30.1, 33.5, and 37.6, respectively (coefficients of variation for the CTs were 3.9%, 5.2%, 4.2%, and 6.4%, respectively). In contrast, genotype 3 was amplified in only 48/64 runs and genotype 2a in only 24/64 runs, presumably because of the lower quantities of these genotypes in the pooled serum. The mean CTs for genotypes 3 and 2a in the pooled control serum were 45.3 and 45.8, respectively (coefficients of variation for the CTs were 5.7% and 5.9%, respectively).

(ii) Amplification cross-reactivity. To assess the possible cross-reactivity of the genotype-specific primer-probe sets, we evaluated the raw data from all samples submitted to the University of Washington molecular virology laboratory for routine testing over a 9-month period, plus the group 1 and group 2 samples described in Materials and Methods. Of the 1,509 total samples with positive genotype amplification (defined as a CT of <45), 987 (65.4%) had amplification with a single genotype-specific primer-probe set, 492 (32.6%) had amplification with two genotype-specific primer-probe sets, 29 (1.9%) had amplification with three genotype-specific primer-probe sets, and 1 sample had amplification with four specific primer-probe sets. Samples with two or more amplifications had a variety of genotype combinations (Table 1). Of samples categorized as genotype 1a, 44.2% also had amplification with the genotype 1b primer-probe set at a CT of <45 (Fig. 1A). However, the 1b amplification was much weaker than the 1a amplification in these samples; in less than 1% were the CTs for the 1a and 1b amplifications within three cycles of each other. Conversely, of samples categorized as genotype 1b, 15.1% had a weaker 1a signal (Fig. 1b). Similar cross-reactivity was seen with the genotype 4 primer-probe set, with 27.7% of the 2b samples, 9.5% of the 2a samples, and 1 of the 12 genotype 6 samples (8.3%) showing weak amplification of genotype 4. If these weaker amplification reactions were due to the presence of a second virus strain with a different genotype, the overall frequency of mixed samples would far exceed that reported in the literature. Thus, the more likely interpretation is that multiple signals can result from false priming or the unexpected presence of matching sequence, especially for the 1b, 1a, and 4 primer-probe sets. This may significantly confound the interpretation of possible mixed-sequence samples.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Frequency of multiple genotype signals


Figure 1
View larger version (39K):
[in this window]
[in a new window]
  		FIG. 1. Scatter distribution of CTs for samples with multiple amplification signals. Each sample is represented along the x axis, ordered from lowest to highest total HCV CT value. At least three data points are shown for each sample: the CT of the total HCV reaction, the CT of the major genotype reaction, and the CT of any additional amplification signal(s) present. (A) Data from samples initially assigned to genotype 1a (n = 514); (B) data from samples from non-1a genotypes (n = 231); (C) corrected CT data from samples initially classified as genotype 1a (n = 296).

Comparison of the real-time RT-PCR HCV genotyping assay to HCV genotyping by RFLP analysis. Initial studies were done to compare genotype results of the real-time assay to results by RFLP analysis. Of the 377 samples tested, 35 (9.3%) had clear amplification with the total HCV primer-probe set but did not show a genotype-specific amplification; these were classified as indeterminate. Of the remaining 342 samples, 251 (73.4%) had identical genotype and subtype results by the two methods (Table 2). Several types of discordant results were seen; 42 (12.3%) were discordant for subtype only, and 42 (12.3%) were discordant for genotype. An additional seven samples had real-time RT-PCR genotyping results suggestive of mixed infection but a single genotype by RFLP. Additional studies were done to clarify the reasons for the discordant results.


View this table:
[in this window]
[in a new window]
  		TABLE 2. Comparison of genotype methods: clinical resultsa

(i) Genotype and subtype discordances. A common discordant result occurred when specimens classified as subtype 1a by the real-time assay were called subtype 1b by RFLP (n = 18). A second set of samples (n = 16) was classified as subtype 1a by the real-time assay but had an indeterminate subtype pattern by RFLP (1a/1b). A total of 42 genotype discordances were seen, 40 of which were classified as genotype 2a by RFLP. The majority of these were classified as 1a (32/40) or 1b (2/40) by the real-time assay. Many of these samples gave subtly atypical RFLP patterns. For 32 of the 34 specimens additional sample was available for sequencing, which in all cases supported the real-time RT-PCR genotyping result. Because of this high apparent error rate in the RFLP 2a genotype subset, all clinical samples typed as 2a by RFLP within the past year (n = 102) were retested with the real-time assay, and a subset of these were sequenced (Table 3). A majority (72%) of the previously typed 2a samples were classified as other genotypes by the real-time assay. Most were found by the real-time assay to be genotype 1a or 1b, but genotypes 4, 5, and 6 were also detected. Among genotype-discordant specimens that were sequenced, most had real-time RT-PCR results in agreement with the sequencing results (31/45). The majority (9/14) of specimens with discrepant results by real-time RT-PCR and sequencing differed only by subtype.


View this table:
[in this window]
[in a new window]
  		TABLE 3. Characterization of RFLP 2a genotype samples

(ii) Indeterminate samples. During initial testing, 37 samples were found that successfully amplified with the total HCV primer-probe but were not detected with any of the genotype-specific primer-probe sets and thus were classified as indeterminate. Over the first year of clinical testing an additional 73 indeterminate samples (of 1,200 total) were identified (a 6.1% rate). Core sequencing for all available samples (n = 87) showed that most indeterminate specimens were either genotype 1b (47.1%) or 1a (25.3%), although other subtypes were also found (14.9% genotype 6, 5.7% genotype 2b, 5.7% genotype 3, and 1.1% genotype 5a). For the 1a, 2b, and 3 groups, these indeterminate samples represented <5% of the total number of samples classified as those genotypes. In contrast, the 41 genotype 1b samples represented 18% of the genotype 1b samples seen and the 13 genotype 6 samples represented more than half of the total genotype 6 samples.

Detection of mixed infections. (i) Possible mixed genotypes as determined by real-time RT-PCR versus sequencing. As noted above, over 30% of specimens showed some degree of amplification with more than one of the genotype-specific primer-probe sets. Although the manufacturer does not provide specific criteria for classifying samples as mixed infections, it has been suggested that the presence of two amplification signals within 3 CTs of each other may indicate the presence of multiple genotypes. During the initial validation and subsequent routine testing, 23 samples (1.4%) were categorized as mixed infections by these criteria (Table 4). Core sequencing did not show the presence of a second sequence for any of the 23 samples, although our sequencing method can detect a second genotype when present at 10 to 20% of total HCV (unpublished). Surprisingly, the genotype amplification with the fastest CT (primary signal) did not always match the genotype determined by sequencing. The most common additional signal not seen by sequencing was 1a (n = 9), followed by 4 (n = 8) and 1b (n = 3).


View this table:
[in this window]
[in a new window]
  		TABLE 4. Mixed-genotype sample results

(ii) Relative amplification of HCV genotypes. Since our previous efficiency estimates were done with a single sample of each genotype, we more closely evaluated the relationship between the total HCV and genotype-specific amplifications among clinical specimens. For each sample, we calculated the difference between the CT of the total HCV signal and that of the genotype-specific signal ({Delta}CT = CTtotal HCV reaction – CTgenotype-specific reaction). Thus, if the efficiency of the total HCV reaction approximated that of the genotype-specific reaction, the mean for the {Delta}CT should approximate zero. Only data from the genotype reaction used to assign the sample genotype were included in this analysis, while mixed-genotype samples, samples with indeterminate genotype, and negative samples were excluded.

The genotype-specific reactions varied widely in their {Delta}CTs (Fig. 2). The mean {Delta}CT for genotype 4 was 2.09, indicating that this amplification occurred more efficiently than the total HCV reaction. The genotype 2a and 2b reactions did not differ significantly from 0 (mean {Delta}CT of –0.68 and –0.04, respectively). Interestingly, the genotype 2b data had a bimodal distribution (mean {Delta}CT of –1.8 and 2.3). Several reactions were significantly less efficient than the total HCV reaction: genotype 1a (mean {Delta}CT of –1.73), genotype 3 (mean {Delta}CT of –3.37), and genotype 1b (mean {Delta}CT of –7.78). Genotype 6 had a mean {Delta}CT of –3.79 but due to low sample numbers could not be tested for a statistically significant difference.


Figure 2
View larger version (21K):
[in this window]
[in a new window]
  		FIG. 2. Scatter distribution of {Delta}CT (CTtotal HCV reaction – CTgenotype-specific reaction) for each sample result with a single genotype amplification, displayed by sample genotype; each point represents one sample. The short horizontal lines represent the mean {Delta}CT for each group. Means for genotypes 1a, 1b, 3, and 4 were all significantly different from zero (P < 0.01), and the genotype 1b mean {Delta}CT was significantly different from all the other means (P < 0.01). Means for genotype 2a and 2b were not significantly different from zero, while there were insufficient data for testing genotype group 5 and 6 means.

(iii) Alternative criteria for classification of mixed samples by real-time RT-PCR. Because the {Delta}CTs for the genotype-specific reactions varied widely, we considered the possibility that the criterion of two amplification signals within 3 CTs might not be appropriate for mixtures of genotypes detected by reactions with greatly different amplification efficiencies. We therefore evaluated whether it might be possible to correct the observed CTs based on the relative efficiencies of each genotype-specific reaction. Among samples with a single amplification of genotype 1b only, an association was noted between the genotype 1b CT and the {Delta}CT such that later genotype CT values corresponded to lower {Delta}CT values (Fig. 3). Similar associations were observed for genotypes 1a and 3 (data not shown). For these genotypes, the regression line defining the relationship between the genotype CT and {Delta}CT was used to calculate a "corrected CT" for samples with multiple signals that met the analysis criteria (Table 5). For genotypes 2a, 2b, and 4, the {Delta}CT remained relatively constant regardless of the genotype CT, so simple shifts based on the mean {Delta}CT were used for genotype CT adjustments. Genotypes 5 and 6 were not evaluated due to low sample numbers.


Figure 3
View larger version (19K):
[in this window]
[in a new window]
  		FIG. 3. Difference between total HCV reaction CT and genotype 1b reaction CT plotted against genotype 1b CT, for samples having no other genotype-specific reactions with a CT of <45 (n = 159). The calculated regression line is shown.


View this table:
[in this window]
[in a new window]
  		TABLE 5. Adjustments to genotype CTs

We then asked whether use of such a corrected CT might be an effective means to assign genotype and detect mixed samples. Using the corrected CT, many samples were classified differently than by the original classification (Fig. 1C; Table 6). Five samples were classified as mixed genotypes by the original method but as a single genotype by the corrected method. Two samples were classified as mixtures using both criteria but not mixtures of the same genotypes. The most significant change was within those originally classified as 1a, where more than half of the group (175 of 296) were called mixed genotypes (1a/1b) using the corrected CTs. Sequencing was done on nine of these samples, and no evidence for mixed sequence was found. Thus, the cross-reactivity between the 1a and 1b reactions prevents the use of corrected CTs in this manner to detect mixed infections.


View this table:
[in this window]
[in a new window]
  		TABLE 6. Comparison of genotype results from two classification schemesa


arrow
DISCUSSION
 
Overall, the real-time RT-PCR HCV genotyping assay performed well compared to results obtained by RFLP and core sequencing. The real-time assay was quick and simple to perform, with results available within 4 to 5 h. The assay's major deficiency was the relatively high rate (6.1%) of indeterminate results. This rate of amplification failure is similar to those reported for other type-specific and real-time assays (20, 22, 25) and similar to that seen for amplifications of non-5' NC regions of the HCV genome (7, 13, 26, 30). For assignment of common genotypes, there was good agreement between the RT-PCR and RFLP methods at the genotype level. Of the samples where results disagreed, most were due to RFLP errors. A small number of mixed-genotype samples seen by the real-time assay could not be confirmed by sequencing. Although tested in limited numbers, the less common genotypes 4 and 6 were frequently miscalled with both methods. Significant difficulties in correctly identifying genotype 4 and 6 samples have also been reported for other methods (5, 9, 33). Currently, use of the real-time assay would require a second method for typing nonamplified samples, mixes, and rare genotypes.

We used the {Delta}CT calculation from a large number of samples to measure the efficiency of each of the genotype-specific amplification reactions. The substantial spread of the {Delta}CT for some of the reactions leads to the hypothesis that within a given genotype there may be sequence polymorphisms leading to better or worse matching with the primers and probes, and thus genotypes with wide spreads in {Delta}CT data may have more base pair variations in the amplification region than those with {Delta}CTs tightly clustered around a single mean. This is especially true for the 1b primer-probe set, where both the spread and overall assay efficiency may reflect multiple mutations in the primer-probe binding sites located within the NS5 region.

A second type of pattern seen with the genotype 2b data was a bimodal distribution (mean {Delta}CT of –1.8 and 2.3). This may indicate two different subpopulations of genotype 2b samples with slightly different sequences within either the 2b or the total HCV amplicon region, leading to different amplification efficiencies within the group. Sequencing studies are currently under way to determine the explanation for these two subgroups within the 2b group.

The amplification reactions were also used to evaluate possible "cross-reactive" amplification by a primer-probe set not specific for the genotype present in the sample. False amplification was frequently observed, especially with the 1b, 1a, and 4 primer-probe sets. Similar "cross-reactive" patterns have been reported with other type-specific genotype amplification methods (11, 13, 15, 17, 19). In the majority of samples, the correct genotype amplification was sufficiently stronger than the cross-reactive amplification so the correct genotype could be determined. Some samples had two amplifications of approximately equal quantity (CT within 3), and these may demonstrate an increased sensitivity of the real-time method to mixed infections.

Mixed-genotype infections in small numbers of samples have been reported in many studies, with higher mixed rates in multiple-exposure groups such as hemophiliacs, patients on chronic hemodialysis, and injection drug users (1, 27, 32). Detection rates also vary depending on the method used, and in studies where multiple methods are used, mixes seen with one method may not be confirmed by a second method (14, 32). Direct sequencing is a relatively insensitive method to detect minor populations, capable of detecting mixes only if the smaller population is at least 10 to 20% of the total, and has been shown to have significant variation in detection in different laboratories (18). The early Line-Probe assay versions fail to distinguish subtype 1a from 1b, thus making them incapable of detecting the most common mixed sample.

Using the real-time assay, we identified a rate of mixed infections somewhat lower than that reported using other genotype-specific PCR assays (21, 32), but unlike our study, these previous studies focused on multiply exposed groups. None of the mixes identified in our study were confirmed by the core sequence assay. However, since the real-time method is quantitatively linear over 6 to 7 logs, it presumably has much greater sensitivity to detect mixes. For example, it might be possible to detect 10 IU/ml of one genotype while detecting 1 x 106 IU/ml of a second genotype, thus detecting a mix at the extremely low 1:100,000 ratio. This contrasts with the current "gold standard" in detection of mixes, the cloning of individual virus sequences, where more than 100,000 clones would have to be sequenced to detect a similar ratio. For the real-time method to fulfill this potential, however, significant improvements will need to be made in the primer-probe specificity.

One unique advantage of the real-time method is the ability to evaluate the efficiency of each genotype-specific amplification. The analysis that we performed clearly demonstrated unequal efficiencies for the different genotype-specific and total HCV amplifications, and this variability creates a major difficulty in the effective use of the CT values. We attempted to overcome this with a mathematical adjustment of the efficiency for inefficient reactions. Unfortunately, adjusting the CT values in this way was not informative, due to the extensive cross-reactivity of certain genotype-specific reactions. Clearly, the real-time assay can be improved by creating primer-probe sets with better specificity and more equal amplification efficiencies.


arrow
ACKNOWLEDGMENTS
 
This work was supported by a grant from Abbott Laboratories.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue N, D3-100, Seattle, WA 98109. Phone: (206) 667-6793. Fax: (206) 667-4411. E-mail: kjerome{at}fhcrc.org. Back

{triangledown} Published ahead of print on 20 September 2006. Back


arrow
REFERENCES
 
    1
  1. Aitken, C., R. McCaw, D. Jardine, S. Bowden, P. Higgs, O. Nguyen, N. Crofts, and M. Hellard. 2004. Change in hepatitis C virus genotype in injecting drug users. J. Med. Virol. 74:543-545.[CrossRef][Medline]
    2. 2
  2. Bukh, J., R. H. Purcell, and R. H. Miller. 1994. Sequence analysis of the core gene of 14 hepatitis C virus genotypes. Proc. Natl. Acad. Sci. USA 91:8239-8243.[Abstract/Free Full Text]
    3. 3
  3. Bullock, G. C., D. E. Bruns, and D. M. Haverstick. 2002. Hepatitis C genotype determination by melting curve analysis with a single set of fluorescent resonance energy transfer probes. Clin. Chem. 48:2147-2154.[Abstract/Free Full Text]
    4. 4
  4. Chayama, K., A. Tsubota, Y. Arase, S. Saitoh, I. Koida, K. Ikeda, T. Matsumoto, M. Kobayashi, S. Iwasaki, S. Koyama, et al. 1993. Genotype subtyping of hepatitis C virus. J. Gastroenterol. Hepatol. 8:150-156.[Medline]
    5. 5
  5. Chinchai, T., J. Labout, S. Noppornpanth, A. Theamboonlers, B. L. Haagmans, A. D. M. E. Osterhaus, and Y. Poovorawan. 2003. Comparative study of different methods to genotype hepatitis C virus type 6 variants. J. Virol. Methods 109:195-201.[CrossRef][Medline]
    6. 6
  6. Choo, Q.-L., G. Kuo, A. J. Weiner, L. R. Overby, D. W. Bradley, and M. Houghton. 1989. Isolation of a cDNA clone derived from a blood-borne non-A, non-B viral hepatitis genome. Science 244:359-362.[Abstract/Free Full Text]
    7. 7
  7. Corbet, S., J. Bukh, A. Heinsen, and A. Fomsgaard. 2003. Hepatitis C virus subtyping by a core-envelope 1-based reverse transcriptase PCR assay with sequencing and its use in determining subtype distribution among Danish patients. J. Clin. Microbiol. 41:1091-1100.[Abstract/Free Full Text]
    8. 8
  8. Davidson, F., P. Simmonds, J. C. Ferguson, L. M. Jarvis, B. C. Dow, E. A. Follett, C. R. Seed, T. Krusius, C. Lin, and G. A. Medgyesi. 1995. Survey of major genotypes and subtypes of hepatitis C virus using RFLP of sequences amplified from the 5' non-coding region. J. Gen. Virol. 76:1197-1204.[Abstract/Free Full Text]
    9. 9
  9. Dev, A., R. McCaw, V. Sundararajan, S. Bowden, and W. Sievert. 2002. Southeast Asian patients with chronic hepatitis C: the impact of novel genotypes and race on treatment outcome. Hepatology 36:1259-1265.[CrossRef][Medline]
    10. 10
  10. Fisher, L. D., and G. Van Belle. 1993. Biostatistics: a methodology for the health sciences, p. 611-613. John Wiley & Sons, Inc., New York, N.Y.
    11. 11
  11. Forns, X., M. D. Maluenda, F. X. Lopez-Labrador, S. Ampurdanes, E. Olmedo, J. Costa, P. Simonds, J. M. Sanchez-Tapias, M. T. Jimenez de Anta, and J. Rodes. 1996. Comparative study of three methods for genotyping hepatitis C strains in samples from Spanish patients. J. Clin. Microbiol. 34:2516-2521.[Abstract]
    12. 12
  12. Fujigaki, H., M. Takemura, K. Takahashi, Y. Yamada, H. Fujii, H. Wada, K. Saito, H. Ohnishi, and M. Seishima. 2004. Genotyping of hepatitis C virus by melting curve analysis with SYBR green I. Ann. Clin. Biochem. 41:130-132.[CrossRef][Medline]
    13. 13
  13. Furione, M., L. Simoncini, M. Gatti, F. Baldanti, M. G. Revello, and G. Gerna. 1999. HCV genotyping by three methods: analysis of discordant results based on sequencing. J. Clin. Virol. 13:121-130.[CrossRef][Medline]
    14. 14
  14. Giannini, C., F. Gianneli, M. Monti, G. Careccia, M. E. Marrocchi, G. Laffi, P. Gentilini, and A. L. Zignego. 1999. Prevalence of mixed infection by different hepatitis C virus genotypes in patients with hepatitis C virus-related chronic liver disease. J. Lab. Clin. Med. 134:68-73.[CrossRef][Medline]
    15. 15
  15. Holland, P. V., J. M. Barrera, M. G. Ercilla, C. F. T. Yoshida, Y. Wang, G. A. B. de Olim, B. Betlach, K. Kuramoto, and H. Okamoto. 1996. Genotyping hepatitis C virus isolates from Spain, Brazil, China, and Macau by a simplified PCR method. J. Clin. Microbiol. 34:2372-2378.[Abstract]
    16. 16
  16. Kato, N., M. Huikata, U. Ootsuyama, M. Nakagawa, S. Ohkoshi, T. Sugimura, and K. Shimotohno. 1990. Molecular cloning of the human hepatitis C virus genome from Japanese patients with non-A, non-B hepatitis. Proc. Natl. Acad. Sci. USA 87:9524-9528.[Abstract/Free Full Text]
    17. 17
  17. Khan, R. U., C. Y. W. Tong, S. Bloom, I. T. Gilmore, C. H. Toh, P. H. Bolton-Maggs, N. J. Beeching, and C. A. Hart. 1997. Evaluation of two simplified methods for genotyping hepatitis C virus. J. Med. Virol. 52:35-41.[CrossRef][Medline]
    18. 18
  18. Laperche, S., F. Lunel, J. Ozopet, et al. 2005. Comparison of hepatitis C virus NS5b and 5' noncoding gene sequencing methods in a multicenter study. J. Clin. Microbiol. 43:733-739.[Abstract/Free Full Text]
    19. 19
  19. Lee, J.-H., W. K. Roth, and S. Zeuzen. 1997. Evaluation and comparison of different hepatitis C virus genotyping and serotyping assays. J. Hepatol. 26:1001-1009.[CrossRef][Medline]
    20. 20
  20. Lindh, M., and C. Hannoun. 2005. Genotyping of hepatitis C virus by Taqman real-time PCR. J. Clin. Virol. 34:108-114.[CrossRef][Medline]
    21. 21
  21. Matsubara, T., R. Sumazaki, K. Shin, Y. Nagai, and H. Takita. 1996. Genotyping of hepatitis C virus: coinfection by multiple genotypes detected in children with chronic posttransfusion hepatitis C. J. Pediatr. Gastroenterol. Nutr. 22:79-84.[CrossRef][Medline]
    22. 22
  22. Ohno, T., M. Mizokami, R.-R. Wu, M. G. Saleh, K.-I. Ohba, E. Orito, M. Mukaide, R. Williams, and J. Y. N. Lau. 1997. New hepatitis C virus (HCV) genotyping system that allows for identification of HCV genotypes 1a, 1b, 2a, 2b, 3a, 3b, 4, 5a, and 6a. J. Clin. Microbiol. 35:201-207.[Abstract]
    23. 23
  23. Okamoto, H., S. Okada, Y. Sugiyama, T. Tanaka, H. Yoshizawa, Y. Miyakawa, and M. Mayumi. 1990. Detection of hepatitis C virus RNA by a two-stage polymerase chain reaction with two pairs of primers deduced from the 5' noncoding region. Jpn. J. Exp. Med. 60:167-177.[Medline]
    24. 24
  24. Okamoto, H., Y. Sugiyama, S. Okada, K. Kurai, Y. Akahane, Y. Sugai, T. Tanaka, K. Sato, F. Tsuda, Y. Miyakawa, and M. Mayumi. 1992. Typing hepatitis C virus by polymerase chain reaction with type-specific primers. J. Gen. Virol. 73:673-679.[Abstract/Free Full Text]
    25. 25
  25. Okamoto, H., S. Kobata, H. Tokita, T. Inoue, G. D. Woodfield, P. V. Holland, B. A. Al-Knawy, O. Uzunalimoglu, Y. Miyakawa, and M. Mayumi. 1996. A second-generation method of genotyping hepatitis C virus by the polymerase chain reaction with sense and antisense primers deduced from the core gene. J. Virol. Methods 57:31-45.[CrossRef][Medline]
    26. 26
  26. Othman, S. B., A. Trabelsi, A. Monnet, N. Bouzgarrou, F. Grattar, A. Beyou, T. Bourlet, and B. Pozzetto. 2004. Evaluation of a prototype HCV NS5b assay for typing strains of hepatitis C virus isolated from Tunisian haemodialysis patients. J. Virol. Methods 119:177-181.[CrossRef][Medline]
    27. 27
  27. Qian, K.-P., S. N. Natov, B. J. G. Pereira, and J. Y.-N. Lau. 2000. Hepatitis C virus mixed genotype infection in patients on haemodialysis. J. Viral Hepat. 7:153-160.[CrossRef][Medline]
    28. 28
  28. Rolfe, K. J., G. J. M. Alexander, T. G. Wreghitt, S. Parmar, H. Jalal, and M. D. Curran. 2005. A real-time Taqman method for hepatitis C virus genotyping. J. Clin. Virol. 34:115-121.[CrossRef][Medline]
    29. 29
  29. Rutledge, R. G., and C. Cote. 2003. Mathematics of quantitative kinetic PCR and the application of standard curves. Nucleic Acids Res. 31:e93.[Abstract/Free Full Text]
    30. 30
  30. Sandres-Saune, K., P. Deny, C. Pasquier, V. Thibaut, G. Duverlie, and J. Izopet. 2003. Determining hepatitis C genotype by analyzing the sequence of the NS5b region. J. Virol. Methods 109:187-193.[CrossRef][Medline]
    31. 31
  31. Schröter, M., B. Zöllner, P. Schäfer, O. Landt, U. Lass, R. Laufs, and H.-H. Feucht. 2002. Genotyping of hepatitis C virus types 1, 2, 3, and 4 by a one-step LightCycler method using three different pairs of hybridization probes. J. Clin. Microbiol. 40:2046-2050.[Abstract/Free Full Text]
    32. 32
  32. Tuveri, R., C. Rothschild, S. Pol, D. Reijasse, T. Persico, C. Gazengel, C. Brechot, and V. Thiers. 1997. Hepatitis C virus genotypes in French haemophiliacs: kinetics and reappraisal of mixed infections. J. Med. Virol. 51:36-41.[CrossRef][Medline]
    33. 33
  33. Zekri, A. R. N., H. M. A. El-Din, A. A. Bahnassy, A. M. R. El-Shehabi, H. El-Leethy, A. Omar, and H. M. Khaled. 2005. TRUGENE sequencing verses INNO-LIPA for subtyping of HCV genotype-4. J. Med. Virol. 75:412-420.[CrossRef][Medline]

Journal of Clinical Microbiology, November 2006, p. 4149-4156, Vol. 44, No. 11
0095-1137/06/$08.00+0     doi:10.1128/JCM.01230-06
Copyright © 2006, American Society for Microbiology. All Rights Reserved.



This article has been cited by other articles:

  • Martro, E., Gonzalez, V., Buckton, A. J., Saludes, V., Fernandez, G., Matas, L., Planas, R., Ausina, V. (2008). Evaluation of a New Assay in Comparison with Reverse Hybridization and Sequencing Methods for Hepatitis C Virus Genotyping Targeting Both 5' Noncoding and Nonstructural 5b Genomic Regions. J. Clin. Microbiol. 46: 192-197 [Abstract] [Full Text]  

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cook, L.
Right arrow Articles by Jerome, K. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Cook, L.
Right arrow Articles by Jerome, K. R.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»