INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Hepatitis C virus (HCV) is an RNA virus with a high rate of genetic mutation (20). As a result, extensive genetic heterogeneity of HCV exists in infected individuals, and HCV isolates are found as either a group of isolates with very closely related genomes, referred to as quasispecies, or genetically distinct groups called "genotypes" (5, 18). At present, 11 types and at least 50 subtypes of HCV have been described (23-25). More than one genotype can be found in the circulations of some HCV-infected patients, particularly in individuals who have received multiple transfusions and intravenous drug users. These are referred to as mixed-genotype infections (13, 26). The rate of HCV mixed-genotype infections is extremely variable in the same group of patients tested by different assays (7). The frequency of distribution of genotypes between patients with single- and mixed-genotype infections was not found to be concordant by the use of current genotyping methods (6). It has been difficult to assess the true prevalence of mixed-genotype infections by currently available assays, including direct DNA sequencing, since they are designed to identify only the HCV genotype dominant in the population (6, 16). As a result, genotypes present at lower proportional levels in a mix could be missed or mistyped (11). Population-based DNA sequencing (i.e., cloning and sequencing of HCV cDNA) is the only reliable method for detection of mixed-genotype infections. However, it is not practical for large cohort studies since it is expensive and time-consuming. Thus, virtually all studies of the clinical significance of HCV genotypes have been based on data from the detection of a single-genotype infection, ignoring the presence of mixed-genotype infections. This may have a significant impact on the interpretation of results from some studies of the biological and clinical differences between HCV genotypes (6, 26). We recently developed a simple and accurate genotyping method, primer-specific and mispair extension analysis (PSMEA) (11). The sensitivity of this assay for detection of HCV mixed-genotype infections is superior to that of direct DNA sequencing. This assay has now been developed as a semiautomated genotyping system (S-PSMEA) that has shown a greater sensitivity and accuracy for detection of human immunodeficiency virus type 1 mixed-genotype infection than our first-generation assay (12). We report here on the sensitivity and reliability of S-PSMEA for detection of HCV mixed-genotype infections by comparing four major currently available genotyping assays, type-specific PCR (T-S PCR), restriction fragment length polymorphism (RFLP) analysis, line probe assay (LiPA), and direct DNA sequencing with S-PSMEA and population-based DNA sequencing. By using S-PSMEA and direct DNA sequencing, the prevalence of HCV mixed-genotype infections was also estimated among HCV-positive blood donors, patients with chronic hepatitis C, and patients with thalassaemia and HCV viremia who had received multiple transfusions in Canada.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Samples. A total of 292 HCV-positive plasma samples were genotyped, of which 152 were from random blood donors who tested positive by both enzyme immunoassay (Ortho Diagnostic Systems Inc., Raritan, N.J.) and HCV RIBA-2 assays (CHIRON, Emeryville, Calif.), 105 were from patients with chronic hepatitis C, and 35 were from thalassemia patients. In addition, 142 HCV RNA-positive samples were selected from 1,280 consecutive HCV RNA-positive patients, including injection drug users, individuals who had received multiple transfusions, and persons with chronic liver disease from northern Italy, for a study on HCV mixed-genotype infections (8; M. Furione, M. Gatti, L. Dossena, C. Zanello, F. Baldanti, and G. Gerna, 5th Int. Meet. Viral Hepatitis C Virus and Related Viruses, Molecular Virology and Pathogenesis, p. 167 (abstr., 1998). The majority (76.9%) of these samples were found to contain more than one genotype according to primer-specific PCR (P-S PCR) methods, but these results could not be confirmed by RFLP analysis or direct DNA sequencing.

cDNA amplifications by PCR. Viral RNA from each plasma sample was purified with the QIAamp Viral RNA kit (QIAGEN Inc., Venlo, The Netherlands). The Pharmacia Biotech First-Strand cDNA synthesis kit was then used for cDNA synthesis. Reverse transcription-PCR was performed with a set of primers (sense primer from positions 302 to 278 [5'-CTCCCCTGTGAGGAACTACTGTCTT-3'] and antisense primer from positions 21 to 1 [5'-GGTGCACGGTCTACGAGACCT-3']) that target highly conserved domains within the 5' untranslated region (5' UR). If a second-round PCR was necessary to provide sufficient cDNA for PSMEA, a pair of nested primers (sense primer from positions 264 to 239 [5'-TCTAGCCATGGCGTTAGTATGAGTGT-3'] and antisense primer from positions 50 to 29 [5'-CTCGCAAGCACCCTATCAGG-3']) was used. The free nucleotides and primers were removed from the PCR products with the QIAquick PCR purification kit (QIAGEN Inc.). These purified PCR products were used for primer extension and automated sequencing analysis (11).

S-PSMEA for HCV genotyping. S-PSMEA for HCV genotyping was performed by a procedure described previously (11, 12). The original assay, PSMEA (12), was improved by substituting the ³²P-labeled primer with a 5'-end Cy 5.5 dye-labeled primer, allowing the assay to be performed with an automated DNA sequencer to detect primer extension. This improvement has led to the development of the assay as a semiautomated system. Primer extension reaction mixtures contained 5 ng of 5'-end Cy 5.5 dye-labeled primer, 5 to 7 ng of PCR product, each required deoxynucleoside triphosphate (dNTP) at a concentration of 20 µM, 0.3 U of Pfu (Pyrococcus furiosus) DNA polymerase, and 2.5 µl of 10× Pfu reaction buffer (Stratagene, La Jolla, Calif.). Primer extensions were performed in a 25-µl reaction volume in a thermocycler (GeneAmp 9600; Perkin-Elmer). Twenty cycles of 94°C for 20 s, 64°C for 20 s, and 72°C for 35 s were performed. One microliter of the primer extension products was mixed with 1 µl of the sequencing stop solution (Visible Genetics, Toronto, Ontario, Canada), and the mixture was electrophoresed on 6% polyacrylamide-8 M urea-Tris-borate-EDTA minigels for 10 min. Extension products were analyzed with automated DNA fragment length polymorphism analysis (FLPA) software in the OpenGene automated DNA sequencing system (Visible Genetics). The principle of primer design for PSMEA has been described previously (11).

RFLP analysis, T-S PCR, and LiPA for HCV genotyping. RFLP and T-S PCR for HCV genotyping were performed as described previously by Furione et al. (8), as modified from the methods of Ponjanpelto et al. (22) and Okamoto et al. (21). LiPA (Inno-LiPA HCV II) was performed by the procedure provided by the manufacturer (Innogenetics N.V., Ghent, Belgium).

Direct and population-based DNA sequencing. Both direct DNA sequencing and population-based sequencing were performed with an automated DNA sequencer (Visible Genetics) and DNA sequencing kits (Visible Genetics). For population-based DNA sequencing, the PCR products were directly cloned into the TOPO TA vector (Invitrogen Corp., Carlsbad, Calif.). The HCV sequence was amplified from positive colonies of Escherichia coli with M13 primers (Pharmacia Biotech). The reamplified PCR products were sequenced by the same procedure used for direct DNA sequencing.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Sensitivity and reliability of S-PSMEA for detection of HCV mixed-genotype infection. PSMEA was based on the unique 3'5' exonuclease proofreading properties of Pfu DNA polymerase (11). It has now been developed as a semiautomated genotyping system with FLPA computer software in an automated DNA sequencing system. Figure 1 illustrates the principle and procedure of S-PSMEA. Briefly, mispair formations and extensions occur during DNA synthesis in a primer extension reaction with an incomplete set of dNTPs when DNA polymerases are used. However, the 3'5' exonuclease activity of Pfu polymerase peaks sharply at its standard polymerization temperature, being extremely discriminative in nucleotide incorporation and proofreading at the initiation step of DNA synthesis. The efficient 3'5' exonuclease proofreading of Pfu can completely prevent primer extension when a mispair (or mispairs) is found at the site of initiation of DNA synthesis (Fig. 1B-1). However, Pfu cannot completely proofread, as it does allow some single mispair formations and extensions at certain nucleotide positions that are several nucleotides away from the primer extension initiation site (Fig. 1B-2 and B-3). Any two or more consecutive mispairs or a single mispair closely followed by another mispair (or mispairs) located downstream of the primer can completely terminate primer extension by Pfu (Fig. 1B-2 and B-3). Taking advantage of the unique property of the Pfu 3'5' exonuclease proofreading activity, nucleotide variations including a single-nucleotide variation, deletion, or insertion can be accurately identified by single PSMEA. In this improved PSMEA, S-PSMEA, a single Cy 5.5 dye-labeled primer extends up to 15 bases, providing a genotype-specific profile on a minigel that can be detected and analyzed with a laser detection system and FLPA software in an automated DNA sequencer (Fig. 1C). Excess primer is used in PSMEA (molar ratio of primer to template of 10:1), so that more than 30% of the total primer could be extended in a 20-cycle primer extension reaction with a thermocycler. Cy 5.5 dye-labeled primers extend, revealing a few peaks with strong signals which are readily quantified with FLPA software, so that the proportion of each genotype in a mix can be measured.

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Principle and procedure of S-PSMEA. (A) Amplification of 5' UR of the HCV genome by PCR. (B) PSMEA reaction with an incomplete set of dNTPs (dCTP and dGTP) and a Cy 5.5 dye-labeled primer. (B-1) Primer extension cannot initiate if a mispair(s) is present at and/or within the 3' end of the primer. (B-2) Single mispair formation and extension may occur at certain nucleotide positions that are more than one nucleotide away from the initiation site of primer extension. (B-3) Two consecutive mispairs or a single mispair closely followed by more than one mispair can completely terminate primer extension. (C) Genotype-specific profiles of the PSMEA reaction determined by FLPA after electrophoresis of PSMEA reaction products. (Top profile) No extension, showing a single peak for an unextended primer, due to a mispair that exists at the nucleotide position immediately adjacent to the 3' end of the primer (see B-1). (Bottom two profiles) Primer extended with the addition of a different number of bases, showing a genotype-specific pattern (also see B-2 and B-3). d, c, and a indicate the position of each detected peak.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Evaluation of the sensitivity of S-PSMEA for detection of low levels of HCV mixed-genotype infections with a defined mixture of genotypes 1b and 2a in different proportions. (A) FLPA graphs showing the genotype-specific profiles and proportion of primer extension in the PSMEA reactions with different proportions of genotypes in the mixtures. The first peak from the left is the unextended primer, and the other two peaks represent extended primer to which different numbers of bases were added. See the report by Hu et al. (11) for details. (B) Quantification analysis of primer extension with different percentages of genotype 2a in the mixture with FLPA software. The proportion of primer extension in agreement with the percentage of the template (genotype 2a) in the mixtures was shown to be 100%. The formula for the calculation was as follows: the percentage of each genotype in the mixture is equal to the [percent primer extension of target genotype/(percent primer extension of target genotype + percent primer extension for each other genotype in the mix)] × 100.

View larger version (19K): [in this window] [in a new window]   FIG. 3.   Detection of mixed-genotype infections (genotypes 1a and 1b) by S-PSMEA. (A) Genotypic profiles of primer extension of genotypes 1a and 1b by PSMEA reactions with universal primer 1BR in presence of different sets of dNTPs. (A-1) No extension with dTTP and dGTP in the presence of genotype 1b alone; (A-2) primer extension profile with dCTP and dGTP in the presence of genotypes 1a and 1b in a mixture; (A-3) no extension with dCTP and dGTP in the presence of genotype 1a alone; (A-4) primer extension profile with dTTP and dGTP in the presence of genotypes 1a and 1b in a mixture. (B) Nucleotide sequences of the templates (genotypes 1a and 1b), primers, and extension products illustrated in panel A. The capital letters CGCGGGGC and GT represent the bases that were extended. The majority of primer extension products ended at a pair with a single mismatch: g in genotype 1b or a in genotype 1a (B-2 and B-4). A small proportion of primers extended with several more bases (gcgccc or ggggg) and ended at a pair with a double mismatch (xx) (B-2 and B-4). The percentage of genotypes 1a and 1b in a mixture is calculated as follows: [percent primer extension of genotype 1a or 1b/percent total primer extension of genotype 1a + genotype 1b (i.e., percent primer extension with G and T + percent primer extension with C and G)] × 100.

Comparison of accuracies of S-PSMEA, RFLP analysis, T-S PCR, and LiPA for genotyping and detection of HCV mixed-genotype infections. For general HCV genotyping, most current genotyping methods are quite reliable (16). However, considerable discordance between those methods has been reported (6, 8; Furione et al., 5th Int. Meet. Viral Hepatitis C and Related Viruses). This may be due to differences in genotype distributions and variations of HCV (23; Furione et al., 5th Int. Meet. Viral Hepatitis C and Related Viruses). To evaluate the accuracy and reliability of widely used genotyping assays and S-PSMEA for general genotyping, a total of 208 samples collected from northern Italy and across Canada were genotyped by direct DNA sequencing, S-PSMEA, RFLP, P-S PCR, and LiPA. Table 3 shows that for isolates of a single genotype or a predominant genotype, the levels of concordance of the results of S-PSMEA, RFLP analysis, P-S PCR, and LiPA, with those of direct DNA sequencing were 98.9, 87.6, 13.6, and 96.7%, respectively. Note that the level of concordance between P-S PCR and direct DNA sequencing was extremely low. This was mainly due to the cross-reactivity between genotype 4 with other genotypes in the P-S PCR (8; Furione et al., 5th Int. Meet. Viral Hepatitis C and Related Viruses). RFLP analysis could accurately identify genotype 4, so that its accuracy was higher than that of P-S PCR. However, in some cases it could not distinguish between genotypes 1a and 1b. S-PSMEA and LiPA were more accurate than the other methods. To compare the accuracies of the genotyping methods for detection of mixed-genotype infections, 20 of the 208 samples defined as containing mixtures of different genotypes by direct DNA sequencing were tested. The results show that the detection rates were 100% (20 of 20 samples), 35.7% (5 of 14), 16.7% (1 of 6), and 14.3% (2 of 14) for S-PSMEA, PS-PCR, LiPA, and RFLP analysis, respectively, demonstrating that S-PSMEA was the most accurate method. Furthermore, with several samples the old PSMEA method (11) and the improved S-PSMEA method were compared; a 100% concordance of the results was observed. In this study, 66 of the 208 samples collected from patients with chronic hepatitis C from Canada were genotyped by LiPA, S-PSMEA, and direct DNA sequencing. The results showed that 6 of the 66 samples were found to contain mixed genotypes by both S-PSMEA and direct DNA sequencing. LiPA detected mixtures of genotypes in only one of the six samples, and three of four samples with mixtures of genotypes of 1a and 1b were found by LiPA to contain genotype 1 isolates, but LiPA provided no subtyping results for these samples.

Prevalence of HCV mixed-genotype infections estimated by S-PSMEA and direct DNA sequencing. Table 4 shows the prevalence of HCV mixed-genotype infections determined by S-PSMEA and direct DNA sequencing in different populations across Canada. As expected, the prevalence of HCV mixed-genotype infections was high (17.1%) in thalassaemia patients who had received multiple transfusions. However, the rate of HCV mixed-genotype infections in patients with chronic hepatitis C was as high as 14.3%, which was not significantly different from that in the thalassaemia patients who had received multiple transfusions. Even in the group of HCV-positive blood donors, the prevalence of mixed-genotype infections was 7.9%, which was much higher than expected. Thirteen (61.9%) of 21 of the mixed-genotype infections in chronic hepatitis C and thalassaemia patients were also detected by direct DNA sequencing. The remaining 38.5% of mixed-genotype infections identified by S-PSMEA could not be detected by direct DNA sequencing, possibly because direct DNA sequencing was not sensitive enough to detect low levels of mixed genotypes, as shown in Table 1 and our previous report (11).

HCV genotype distribution in patients with single- and mixed-genotype infections. Table 5 shows the genotype distribution of HCV in blood donors and thalassemia patients infected with single or multiple HCV genotypes in Canada as determined by S-PSMEA. Among the samples with single-genotype infections, genotype 1a was the predominant genotype in blood donors (41.5%) and thalassemia patients who had received multiple transfusions (51.7%). Among the samples with mixed-genotype infections, genotype 1a was found in 71.4% of the blood donors and 100% of the thalassemia patients. Genotypes 1b and 2b were the next commonly found genotypes in samples either with single-genotype infections or with mixed-genotype infections. It was apparent that the distribution of genotypes in samples with HCV single- and mixed-genotype infections determined by S-PSMEA was concordant for both blood donors and thalassaemia patients. Samples were collected from blood donors in western and eastern Canada, with a higher prevalence of genotype 6a frequently found in the samples from western Canada. The samples from thalassemia patients were obtained from patients in Toronto, in eastern Canada; thus, no isolates of genotype 6a were found in this group of patients.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study we have demonstrated that PSMEA is more reliable than other currently available genotyping methods for detection of HCV mixed-genotype infections. This has allowed us to more accurately estimate the prevalence of mixed-genotype infections in different populations, which has important implications for clinical and epidemiological studies of HCV genotypes (3, 17, 19).

The accuracies of four currently available genotyping methods for detection of HCV mixed-genotype infections compared with those of S-PSMEA and population-based DNA sequencing were reevaluated in this study. It appears from our data that S-PSMEA is a more sensitive method than T-S PCR, RFLP analysis, and LiPA for detection of HCV mixed-genotype infections. Although direct DNA sequencing remains the best among the currently available genotyping assays, only about 60% of HCV mixed-genotype infections can be detected by that method. In light of these findings, results by these methods that indicate HCV mixed-genotype infections should be interpreted with caution.

Genotypes 1a and 1b are the most prevalent in most parts of the world (23). On the basis of our results, 40 to 75% of mixed-genotype 1a and 1b infections can be missed or mistyped by these currently available genotyping methods. The true prevalence of HCV mixed-genotype infections, particularly, mixed-genotype 1a and 1b infections, should be higher than that estimated previously by the widely used genotyping methods with samples from blood donors and patients with chronic hepatitis C (4). In addition, it is difficult for these methods to identify more than two genotypes in a mixture. Extensive cloning and sequencing of a large number of clones from each of these isolates are required in order to determine whether more than one genotype is present simultaneously. Unfortunately, use of this method for clinical studies is impractical. S-PSMEA is an alternative for accurate detection of HCV mixed-genotype infections including multiple-genotype infections in large cohort studies.

It is not surprising that a high frequency of mixed-genotype infections is found in persons with hemophilia since many had multiple exposures to HCV from long-term use of older coagulation factor concentrates before heat treatment and anti-HCV testing was implemented (13, 26). In this study, the prevalence of HCV mixed-genotype infections among blood donors and patients with chronic hepatitis determined by S-PSMEA was found to be higher than that reported previously (4, 16). This is because low levels of a genotype present in a mixture of genotypes can be detected by S-PSMEA. On the basis of epidemiological studies, mixed-genotype infections in HCV-positive patients should be common because injection drug use currently accounts for 60% of cases of HCV transmission and has accounted for a substantial proportion of HCV infections in past decades (1, 2). Multiple exposures may have led to episodes of reinfection with different genotypes for many chronic injection drug users (9). It is noted in this study that the rate of mixed-genotype infections in thalassemia patients who have received multiple transfusions was not as high as expected. Our explanation is that these patients are young (i.e., age, 22.3 ± 10.7 years), so that the frequency of their exposure to HCV has been dramatically decreased because the screening tests for HCV were introduced in 1990.

Reliable detection of HCV mixed-genotype infections may have broad applications for studies of the clinical significance of HCV genotypes and epidemiological mysteries: (i) accurate detection of mixed-genotype infections is required for study of the frequency of HCV reinfection and reactivation (14). (ii) Reliable detection of mixed-genotype infections is crucial to gain a precise understanding of the mechanism of genotype changes over time in patients with conditions such as hemophilia and thalassemia who have received multiple transfusions (26). (iii) A high frequency of mixed-genotype 1a and 1b infections may have a significant impact on clinical studies of whether genotype 1b is less sensitive to interferon treatment and is associated with a more severe form of liver disease than the form of liver disease associated with other genotypes (10). (iv) Analysis of samples from individuals with mixed-genotype infections can facilitate studies of the clinical and biological differences among HCV genotypes in which confounding host factors would be completely eliminated (10). (v) The prevalence of HCV mixed-genotype infections can be used as a marker for monitoring the presence of risk factors in a population. (vi) Accurate detection of mixed-genotype infection provides a tool that can be used to address whether multiple episodes of reinfection with different genotypes would increase the likelihood of acquiring highly pathogenic variants (or genotype), resulting in more severe hepatitis when compared with that from a primary single-genotype infection (15; Y. W. Hu, E. Balaskas, G. Kessler, V. Scalia, A. Giulivi, G. Sher, and P. Gill, 5th Int. Meet. Viral Hepatitis C Virus and Related Viruses, Molecular Virology and Pathogenesis, p. 179 (abstr., 1998).